activerecord 6.1.0.rc1 on Ruby Gems

Active Support

Calling iso8601 on negative durations retains the negative sign on individual
digits instead of prepending it.
This change is required so we can interoperate with PostgreSQL, which prefers
negative signs for each component.
Compatibility with other iso8601 parsers which support leading negatives as well
as negatives per component is still retained.
Before:
```
(-1.year - 1.day).iso8601
# => "-P1Y1D"
```
After:
```
(-1.year - 1.day).iso8601
# => "P-1Y-1D"
```
Vipul A M
Remove deprecated ActiveSupport::Notifications::Instrumenter#end=.
Rafael Mendonça França
Deprecate ActiveSupport::Multibyte::Unicode.default_normalization_form.
Rafael Mendonça França
Remove deprecated ActiveSupport::Multibyte::Unicode.pack_graphemes,
ActiveSupport::Multibyte::Unicode.unpack_graphemes,
ActiveSupport::Multibyte::Unicode.normalize,
ActiveSupport::Multibyte::Unicode.downcase,
ActiveSupport::Multibyte::Unicode.upcase and ActiveSupport::Multibyte::Unicode.swapcase.
Rafael Mendonça França
Remove deprecated ActiveSupport::Multibyte::Chars#consumes? and ActiveSupport::Multibyte::Chars#normalize.
Rafael Mendonça França
Remove deprecated file active_support/core_ext/range/include_range.
Rafael Mendonça França
Remove deprecated file active_support/core_ext/hash/transform_values.
Rafael Mendonça França
Remove deprecated file active_support/core_ext/hash/compact.
Rafael Mendonça França
Remove deprecated file active_support/core_ext/array/prepend_and_append.
Rafael Mendonça França
Remove deprecated file active_support/core_ext/numeric/inquiry.
Rafael Mendonça França
Remove deprecated file active_support/core_ext/module/reachable.
Rafael Mendonça França
Remove deprecated Module#parent_name, Module#parent and Module#parents.
Rafael Mendonça França
Remove deprecated ActiveSupport::LoggerThreadSafeLevel#after_initialize.
Rafael Mendonça França
Remove deprecated LoggerSilence constant.
Rafael Mendonça França
Remove deprecated fallback to I18n.default_local when config.i18n.fallbacks is empty.
Rafael Mendonça França
Remove entries from local cache on RedisCacheStore#delete_matched
Fixes #38627
ojab
Speed up ActiveSupport::SecurityUtils.fixed_length_secure_compare by using
OpenSSL.fixed_length_secure_compare, if available.
Nate Matykiewicz

ActiveSupport::Cache::MemCacheStore now checks ENV["MEMCACHE_SERVERS"] before falling back to "localhost:11211" if configured without any addresses.

config.cache_store = :mem_cache_store

# is now equivalent to

config.cache_store = :mem_cache_store, ENV["MEMCACHE_SERVERS"] || "localhost:11211"

# instead of

config.cache_store = :mem_cache_store, "localhost:11211" # ignores ENV["MEMCACHE_SERVERS"]

Sam Bostock

ActiveSupport::Subscriber#attach_to now accepts an inherit_all: argument. When set to true,
it allows a subscriber to receive events for methods defined in the subscriber's ancestor class(es).

class ActionControllerSubscriber < ActiveSupport::Subscriber
  attach_to :action_controller

  def start_processing(event)
    info "Processing by #{event.payload[:controller]}##{event.payload[:action]} as #{format}"
  end

  def redirect_to(event)
    info { "Redirected to #{event.payload[:location]}" }
  end
end

# We detach ActionControllerSubscriber from the :action_controller namespace so that our CustomActionControllerSubscriber
# can provide its own instrumentation for certain events in the namespace
ActionControllerSubscriber.detach_from(:action_controller)

class CustomActionControllerSubscriber < ActionControllerSubscriber
  attach_to :action_controller, inherit_all: true

  def start_processing(event)
    info "A custom response to start_processing events"
  end

  # => CustomActionControllerSubscriber will process events for "start_processing.action_controller" notifications
  # using its own #start_processing implementation, while retaining ActionControllerSubscriber's instrumentation
  # for "redirect_to.action_controller" notifications
end

Adrianna Chang

Allow the digest class used to generate non-sensitive digests to be configured with config.active_support.hash_digest_class.
config.active_support.use_sha1_digests is deprecated in favour of config.active_support.hash_digest_class = ::Digest::SHA1.
Dirkjan Bussink
Fix bug to make memcached write_entry expire correctly with unless_exist
Jye Lee
Add ActiveSupport::Duration conversion methods
in_seconds, in_minutes, in_hours, in_days, in_weeks, in_months, and in_years return the respective duration covered.
Jason York
Fixed issue in ActiveSupport::Cache::RedisCacheStore not passing options
to read_multi causing fetch_multi to not work properly
Rajesh Sharma
Fixed issue in ActiveSupport::Cache::MemCacheStore which caused duplicate compression,
and caused the provided compression_threshold to not be respected.
Max Gurewitz
Prevent RedisCacheStore and MemCacheStore from performing compression
when reading entries written with raw: true.
Max Gurewitz
URI.parser is deprecated and will be removed in Rails 6.2. Use
URI::DEFAULT_PARSER instead.
Jean Boussier
require_dependency has been documented to be obsolete in :zeitwerk
mode. The method is not deprecated as such (yet), but applications are
encouraged to not use it.
In :zeitwerk mode, semantics match Ruby's and you do not need to be
defensive with load order. Just refer to classes and modules normally. If
the constant name is dynamic, camelize if needed, and constantize.
Xavier Noria
Add 3rd person aliases of Symbol#start_with? and Symbol#end_with?.
```
:foo.starts_with?("f") # => true
:foo.ends_with?("o")   # => true
```
Ryuta Kamizono
Add override of unary plus for ActiveSupport::Duration.
+ 1.second is now identical to +1.second to prevent errors
where a seemingly innocent change of formatting leads to a change in the code behavior.
Before:
```
+1.second.class
# => ActiveSupport::Duration
(+ 1.second).class
# => Integer
```
After:
```
+1.second.class
# => ActiveSupport::Duration
(+ 1.second).class
# => ActiveSupport::Duration
```
Fixes #39079.
Roman Kushnir

Add subsec to ActiveSupport::TimeWithZone#inspect.

Before:

Time.at(1498099140).in_time_zone.inspect
# => "Thu, 22 Jun 2017 02:39:00 UTC +00:00"
Time.at(1498099140, 123456780, :nsec).in_time_zone.inspect
# => "Thu, 22 Jun 2017 02:39:00 UTC +00:00"
Time.at(1498099140 + Rational("1/3")).in_time_zone.inspect
# => "Thu, 22 Jun 2017 02:39:00 UTC +00:00"

After:

Time.at(1498099140).in_time_zone.inspect
# => "Thu, 22 Jun 2017 02:39:00.000000000 UTC +00:00"
Time.at(1498099140, 123456780, :nsec).in_time_zone.inspect
# => "Thu, 22 Jun 2017 02:39:00.123456780 UTC +00:00"
Time.at(1498099140 + Rational("1/3")).in_time_zone.inspect
# => "Thu, 22 Jun 2017 02:39:00.333333333 UTC +00:00"

akinomaeni

Calling ActiveSupport::TaggedLogging#tagged without a block now returns a tagged logger.
```
logger.tagged("BCX").info("Funky time!") # => [BCX] Funky time!
```
Eugene Kenny
Align Range#cover? extension behavior with Ruby behavior for backwards ranges.
(1..10).cover?(5..3) now returns false, as it does in plain Ruby.
Also update #include? and #=== behavior to match.
Michael Groeneman
Update to TZInfo v2.0.0.
This changes the output of ActiveSupport::TimeZone.utc_to_local, but
can be controlled with the
ActiveSupport.utc_to_local_returns_utc_offset_times config.
New Rails 6.1 apps have it enabled by default, existing apps can upgrade
via the config in config/initializers/new_framework_defaults_6_1.rb
See the utc_to_local_returns_utc_offset_times documentation for details.
Phil Ross, Jared Beck
Add Date and Time #yesterday? and #tomorrow? alongside #today?.
Aliased to #prev_day? and #next_day? to match the existing #prev/next_day methods.
Jatin Dhankhar
Add Enumerable#pick to complement ActiveRecord::Relation#pick.
Eugene Kenny

[Breaking change] ActiveSupport::Callbacks#halted_callback_hook now receive a 2nd argument:

ActiveSupport::Callbacks#halted_callback_hook now receive the name of the callback
being halted as second argument.
This change will allow you to differentiate which callbacks halted the chain
and act accordingly.

  class Book < ApplicationRecord
    before_save { throw(:abort) }
    before_create { throw(:abort) }

    def halted_callback_hook(filter, callback_name)
      Rails.logger.info("Book couldn't be #{callback_name}d")
    end

    Book.create # => "Book couldn't be created"
    book.save # => "Book couldn't be saved"
  end

Edouard Chin

Support prepend with ActiveSupport::Concern.
Allows a module with extend ActiveSupport::Concern to be prepended.
```
module Imposter
  extend ActiveSupport::Concern

  # Same as `included`, except only run when prepended.
  prepended do
  end
end

class Person
  prepend Imposter
end
```
Class methods are prepended to the base class, concerning is also
updated: concerning :Imposter, prepend: true do.
Jason Karns, Elia Schito
Deprecate using Range#include? method to check the inclusion of a value
in a date time range. It is recommended to use Range#cover? method
instead of Range#include? to check the inclusion of a value
in a date time range.
Vishal Telangre

Support added for a round_mode parameter, in all number helpers. (See: BigDecimal::mode.)

number_to_currency(1234567890.50, precision: 0, round_mode: :half_down) # => "$1,234,567,890"
number_to_percentage(302.24398923423, precision: 5, round_mode: :down) # => "302.24398%"
number_to_rounded(389.32314, precision: 0, round_mode: :ceil) # => "390"
number_to_human_size(483989, precision: 2, round_mode: :up) # => "480 KB"
number_to_human(489939, precision: 2, round_mode: :floor) # => "480 Thousand"

485000.to_s(:human, precision: 2, round_mode: :half_even) # => "480 Thousand"

Tom Lord

Array#to_sentence no longer returns a frozen string.

Before:

['one', 'two'].to_sentence.frozen?
# => true

After:

['one', 'two'].to_sentence.frozen?
# => false

Nicolas Dular

When an instance of ActiveSupport::Duration is converted to an iso8601 duration string, if weeks are mixed with date parts, the week part will be converted to days.
This keeps the parser and serializer on the same page.

duration = ActiveSupport::Duration.build(1000000)
# 1 week, 4 days, 13 hours, 46 minutes, and 40.0 seconds

duration_iso = duration.iso8601
# P11DT13H46M40S

ActiveSupport::Duration.parse(duration_iso)
# 11 days, 13 hours, 46 minutes, and 40 seconds

duration = ActiveSupport::Duration.build(604800)
# 1 week

duration_iso = duration.iso8601
# P1W

ActiveSupport::Duration.parse(duration_iso)
# 1 week

Abhishek Sarkar

Add block support to ActiveSupport::Testing::TimeHelpers#travel_back.
Tim Masliuchenko
Update ActiveSupport::Messages::Metadata#fresh? to work for cookies with expiry set when
ActiveSupport.parse_json_times = true.
Christian Gregg
Support symbolic links for content_path in ActiveSupport::EncryptedFile.
Takumi Shotoku
Improve Range#===, Range#include?, and Range#cover? to work with beginless (startless)
and endless range targets.
Allen Hsu, Andrew Hodgkinson
Don't use Process#clock_gettime(CLOCK_THREAD_CPUTIME_ID) on Solaris.
Iain Beeston

Prevent ActiveSupport::Duration.build(value) from creating instances of
ActiveSupport::Duration unless value is of type Numeric.

Addresses the errant set of behaviours described in #37012 where
ActiveSupport::Duration comparisons would fail confusingly
or return unexpected results when comparing durations built from instances of String.

Before:

small_duration_from_string = ActiveSupport::Duration.build('9')
large_duration_from_string = ActiveSupport::Duration.build('100000000000000')
small_duration_from_int = ActiveSupport::Duration.build(9)

large_duration_from_string > small_duration_from_string
# => false

small_duration_from_string == small_duration_from_int
# => false

small_duration_from_int < large_duration_from_string
# => ArgumentError (comparison of ActiveSupport::Duration::Scalar with ActiveSupport::Duration failed)

large_duration_from_string > small_duration_from_int
# => ArgumentError (comparison of String with ActiveSupport::Duration failed)

After:

small_duration_from_string = ActiveSupport::Duration.build('9')
# => TypeError (can't build an ActiveSupport::Duration from a String)

Alexei Emam

Add ActiveSupport::Cache::Store#delete_multi method to delete multiple keys from the cache store.
Peter Zhu
Support multiple arguments in HashWithIndifferentAccess for merge and update methods, to
follow Ruby 2.6 addition.
Wojciech Wnętrzak
Allow initializing thread_mattr_* attributes via :default option.
```
class Scraper
  thread_mattr_reader :client, default: Api::Client.new
end
```
Guilherme Mansur
Add compact_blank for those times when you want to remove #blank? values from
an Enumerable (also compact_blank! on Hash, Array, ActionController::Parameters).
Dana Sherson

Make ActiveSupport::Logger Fiber-safe.

Use Fiber.current.__id__ in ActiveSupport::Logger#local_level= in order
to make log level local to Ruby Fibers in addition to Threads.

Example:

logger = ActiveSupport::Logger.new(STDOUT)
logger.level = 1
puts "Main is debug? #{logger.debug?}"

Fiber.new {
  logger.local_level = 0
  puts "Thread is debug? #{logger.debug?}"
}.resume

puts "Main is debug? #{logger.debug?}"

Before:

Main is debug? false
Thread is debug? true
Main is debug? true

After:

Main is debug? false
Thread is debug? true
Main is debug? false

Fixes #36752.

Alexander Varnin

Allow the on_rotation proc used when decrypting/verifying a message to be
passed at the constructor level.

Before:

crypt = ActiveSupport::MessageEncryptor.new('long_secret')
crypt.decrypt_and_verify(encrypted_message, on_rotation: proc { ... })
crypt.decrypt_and_verify(another_encrypted_message, on_rotation: proc { ... })

After:

crypt = ActiveSupport::MessageEncryptor.new('long_secret', on_rotation: proc { ... })
crypt.decrypt_and_verify(encrypted_message)
crypt.decrypt_and_verify(another_encrypted_message)

Edouard Chin

delegate_missing_to would raise a DelegationError if the object
delegated to was nil. Now the allow_nil option has been added to enable
the user to specify they want nil returned in this case.
Matthew Tanous
truncate would return the original string if it was too short to be truncated
and a frozen string if it were long enough to be truncated. Now truncate will
consistently return an unfrozen string regardless. This behavior is consistent
with gsub and strip.
Before:
```
'foobar'.truncate(5).frozen?
# => true
'foobar'.truncate(6).frozen?
# => false
```
After:
```
'foobar'.truncate(5).frozen?
# => false
'foobar'.truncate(6).frozen?
# => false
```
Jordan Thomas

Active Model

Pass in base instead of base_class to Error.human_attribute_name
This is useful in cases where the human_attribute_name method depends
on other attributes' values of the class under validation to derive what the
attribute name should be.
Filipe Sabella
Deprecate marshalling load from legacy attributes format.
Ryuta Kamizono

*_previously_changed? accepts :from and :to keyword arguments like *_changed?.

topic.update!(status: :archived)
topic.status_previously_changed?(from: "active", to: "archived")
# => true

George Claghorn

Raise FrozenError when trying to write attributes that aren't backed by the database on an object that is frozen:

class Animal
  include ActiveModel::Attributes
  attribute :age
end

animal = Animal.new
animal.freeze
animal.age = 25 # => FrozenError, "can't modify a frozen Animal"

Josh Brody

Add *_previously_was attribute methods when dirty tracking. Example:

pirate.update(catchphrase: "Ahoy!")
pirate.previous_changes["catchphrase"] # => ["Thar She Blows!", "Ahoy!"]
pirate.catchphrase_previously_was # => "Thar She Blows!"

DHH

Encapsulate each validation error as an Error object.
The ActiveModel’s errors collection is now an array of these Error
objects, instead of messages/details hash.
For each of these Error object, its message and full_message methods
are for generating error messages. Its details method would return error’s
extra parameters, found in the original details hash.
The change tries its best at maintaining backward compatibility, however
some edge cases won’t be covered, like errors#first will return ActiveModel::Error and manipulating
errors.messages and errors.details hashes directly will have no effect. Moving forward,
please convert those direct manipulations to use provided API methods instead.
The list of deprecated methods and their planned future behavioral changes at the next major release are:
- errors#slice! will be removed.
- errors#each with the key, value two-arguments block will stop working, while the error single-argument block would return Error object.
- errors#values will be removed.
- errors#keys will be removed.
- errors#to_xml will be removed.
- errors#to_h will be removed, and can be replaced with errors#to_hash.
- Manipulating errors itself as a hash will have no effect (e.g. errors[:foo] = 'bar').
- Manipulating the hash returned by errors#messages (e.g. errors.messages[:foo] = 'bar') will have no effect.
- Manipulating the hash returned by errors#details (e.g. errors.details[:foo].clear) will have no effect.
lulalala

Active Record

Add connected_to_many API.
This API allows applications to connect to multiple databases at once without switching all of them or implementing a deeply nested stack.
Before:
AnimalsRecord.connected_to(role: :reading) do
MealsRecord.connected_to(role: :reading) do
Dog.first # read from animals replica
Dinner.first # read from meals replica
Person.first # read from primary writer
end
end
After:
ActiveRecord::Base.connected_to_many([AnimalsRecord, MealsRecord], role: :reading) do
Dog.first # read from animals replica
Dinner.first # read from meals replica
Person.first # read from primary writer
end
Eileen M. Uchitelle, John Crepezzi
Add option to raise or log for ActiveRecord::StrictLoadingViolationError.
Some applications may not want to raise an error in production if using strict_loading. This would allow an application to set strict loading to log for the production environment while still raising in development and test environments.
Set config.active_record.action_on_strict_loading_violation to :log errors instead of raising.
Eileen M. Uchitelle
Allow the inverse of a has_one association that was previously autosaved to be loaded.
Fixes #34255.
Steven Weber
Optimise the length of index names for polymorphic references by using the reference name rather than the type and id column names.
Because the default behaviour when adding an index with multiple columns is to use all column names in the index name, this could frequently lead to overly long index names for polymorphic references which would fail the migration if it exceeded the database limit.
This change reduces the chance of that happening by using the reference name, e.g. index_my_table_on_my_reference.
Fixes #38655.
Luke Redpath
MySQL: Uniqueness validator now respects default database collation,
no longer enforce case sensitive comparison by default.
Ryuta Kamizono
Remove deprecated methods from ActiveRecord::ConnectionAdapters::DatabaseLimits.
column_name_length
table_name_length
columns_per_table
indexes_per_table
columns_per_multicolumn_index
sql_query_length
joins_per_query
Rafael Mendonça França
Remove deprecated ActiveRecord::ConnectionAdapters::AbstractAdapter#supports_multi_insert?.
Rafael Mendonça França
Remove deprecated ActiveRecord::ConnectionAdapters::AbstractAdapter#supports_foreign_keys_in_create?.
Rafael Mendonça França
Remove deprecated ActiveRecord::ConnectionAdapters::PostgreSQLAdapter#supports_ranges?.
Rafael Mendonça França
Remove deprecated ActiveRecord::Base#update_attributes and ActiveRecord::Base#update_attributes!.
Rafael Mendonça França
Remove deprecated migrations_path argument in ActiveRecord::ConnectionAdapter::SchemaStatements#assume_migrated_upto_version.
Rafael Mendonça França
Remove deprecated config.active_record.sqlite3.represent_boolean_as_integer.
Rafael Mendonça França

relation.create does no longer leak scope to class level querying methods
in initialization block and callbacks.

Before:

User.where(name: "John").create do |john|
  User.find_by(name: "David") # => nil
end

After:

User.where(name: "John").create do |john|
  User.find_by(name: "David") # => #<User name: "David", ...>
end

Ryuta Kamizono

Named scope chain does no longer leak scope to class level querying methods.

class class User < ActiveRecord::Base
  scope :david, -> { User.where(name: "David") }
end

Before:

User.where(name: "John").david
# SELECT * FROM users WHERE name = 'John' AND name = 'David'

After:

User.where(name: "John").david
# SELECT * FROM users WHERE name = 'David'

Ryuta Kamizono

Remove deprecated methods from ActiveRecord::DatabaseConfigurations.
fetch
each
first
values
[]=
Rafael Mendonça França

where.not now generates NAND predicates instead of NOR.

Before:

 User.where.not(name: "Jon", role: "admin")
 # SELECT * FROM users WHERE name != 'Jon' AND role != 'admin'

After:

 User.where.not(name: "Jon", role: "admin")
 # SELECT * FROM users WHERE NOT (name == 'Jon' AND role == 'admin')

Rafael Mendonça França

Remove deprecated ActiveRecord::Result#to_hash method.
Rafael Mendonça França
Deprecate ActiveRecord::Base.allow_unsafe_raw_sql.
Rafael Mendonça França
Remove deprecated support for using unsafe raw SQL in ActiveRecord::Relation methods.
Rafael Mendonça França
Allow users to silence the "Rails couldn't infer whether you are using multiple databases..."
message using config.active_record.suppress_multiple_database_warning.
Omri Gabay
Connections can be granularly switched for abstract classes when connected_to is called.
This change allows connected_to to switch a role and/or shard for a single abstract class instead of all classes globally. Applications that want to use the new feature need to set config.active_record.legacy_connection_handling to false in their application configuration.
Example usage:
Given an application we have a User model that inherits from ApplicationRecord and a Dog model that inherits from AnimalsRecord. AnimalsRecord and ApplicationRecord have writing and reading connections as well as shard default, one, and two.
```
ActiveRecord::Base.connected_to(role: :reading) do
  User.first # reads from default replica
  Dog.first # reads from default replica

  AnimalsRecord.connected_to(role: :writing, shard: :one) do
    User.first # reads from default replica
    Dog.first # reads from shard one primary
  end

  User.first # reads from default replica
  Dog.first # reads from default replica

  ApplicationRecord.connected_to(role: :writing, shard: :two) do
    User.first # reads from shard two primary
    Dog.first # reads from default replica
  end
end
```
Eileen M. Uchitelle, John Crepezzi
Allow double-dash comment syntax when querying read-only databases
James Adam
Add values_at method.
Returns an array containing the values associated with the given methods.
```
topic = Topic.first
topic.values_at(:title, :author_name)
# => ["Budget", "Jason"]
```
Similar to Hash#values_at but on an Active Record instance.
Guillaume Briday
Fix read_attribute_before_type_cast to consider attribute aliases.
Marcelo Lauxen

Support passing record to uniqueness validator :conditions callable:

class Article < ApplicationRecord
  validates_uniqueness_of :title, conditions: ->(article) {
    published_at = article.published_at
    where(published_at: published_at.beginning_of_year..published_at.end_of_year)
  }
end

Eliot Sykes

BatchEnumerator#update_all and BatchEnumerator#delete_all now return the
total number of rows affected, just like their non-batched counterparts.
```
Person.in_batches.update_all("first_name = 'Eugene'") # => 42
Person.in_batches.delete_all # => 42
```
Fixes #40287.
Eugene Kenny

Add support for PostgreSQL interval data type with conversion to
ActiveSupport::Duration when loading records from database and
serialization to ISO 8601 formatted duration string on save.
Add support to define a column in migrations and get it in a schema dump.
Optional column precision is supported.

To use this in 6.1, you need to place the next string to your model file:

attribute :duration, :interval

To keep old behavior until 6.2 is released:

attribute :duration, :string

Example:

create_table :events do |t|
  t.string   :name
  t.interval :duration
end

class Event < ApplicationRecord
  attribute :duration, :interval
end

Event.create!(name: 'Rock Fest', duration: 2.days)
Event.last.duration # => 2 days
Event.last.duration.iso8601 # => "P2D"
Event.new(duration: 'P1DT12H3S').duration # => 1 day, 12 hours, and 3 seconds
Event.new(duration: '1 day') # Unknown value will be ignored and NULL will be written to database

Andrey Novikov

Allow associations supporting the dependent: key to take dependent: :destroy_async.
```
class Account < ActiveRecord::Base
    belongs_to :supplier, dependent: :destroy_async
end
```
:destroy_async will enqueue a job to destroy associated records in the background.
DHH, George Claghorn, Cory Gwin, Rafael Mendonça França, Adrianna Chang
Add SKIP_TEST_DATABASE environment variable to disable modifying the test database when rails db:create and rails db:drop are called.
Jason Schweier
connects_to can only be called on ActiveRecord::Base or abstract classes.
Ensure that connects_to can only be called from ActiveRecord::Base or abstract classes. This protects the application from opening duplicate or too many connections.
Eileen M. Uchitelle, John Crepezzi
All connection adapters execute now raises ActiveRecord::ConnectionNotEstablished rather than
ActiveRecord::StatementInvalid when they encounter a connection error.
Jean Boussier
Mysql2Adapter#quote_string now raises ActiveRecord::ConnectionNotEstablished rather than
ActiveRecord::StatementInvalid when it can't connect to the MySQL server.
Jean Boussier
Add support for check constraints that are NOT VALID via validate: false (PostgreSQL-only).
Alex Robbin
Ensure the default configuration is considered primary or first for an environment
If a multiple database application provides a configuration named primary, that will be treated as default. In applications that do not have a primary entry, the default database configuration will be the first configuration for an environment.
Eileen M. Uchitelle

Allow where references association names as joined table name aliases.

class Comment < ActiveRecord::Base
  enum label: [:default, :child]
  has_many :children, class_name: "Comment", foreign_key: :parent_id
end

# ... FROM comments LEFT OUTER JOIN comments children ON ... WHERE children.label = 1
Comment.includes(:children).where("children.label": "child")

Ryuta Kamizono

Support storing demodulized class name for polymorphic type.
Before Rails 6.1, storing demodulized class name is supported only for STI type
by store_full_sti_class class attribute.
Now store_full_class_name class attribute can handle both STI and polymorphic types.
Ryuta Kamizono
Deprecate rails db:structure:{load, dump} tasks and extend
rails db:schema:{load, dump} tasks to work with either :ruby or :sql format,
depending on config.active_record.schema_format configuration value.
fatkodima

Respect the select values for eager loading.

post = Post.select("UPPER(title) AS title").first
post.title # => "WELCOME TO THE WEBLOG"
post.body  # => ActiveModel::MissingAttributeError

# Rails 6.0 (ignore the `select` values)
post = Post.select("UPPER(title) AS title").eager_load(:comments).first
post.title # => "Welcome to the weblog"
post.body  # => "Such a lovely day"

# Rails 6.1 (respect the `select` values)
post = Post.select("UPPER(title) AS title").eager_load(:comments).first
post.title # => "WELCOME TO THE WEBLOG"
post.body  # => ActiveModel::MissingAttributeError

Ryuta Kamizono

Allow attribute's default to be configured but keeping its own type.

class Post < ActiveRecord::Base
  attribute :written_at, default: -> { Time.now.utc }
end

# Rails 6.0
Post.type_for_attribute(:written_at) # => #<Type::Value ... precision: nil, ...>

# Rails 6.1
Post.type_for_attribute(:written_at) # => #<Type::DateTime ... precision: 6, ...>

Ryuta Kamizono

Allow default to be configured for Enum.

class Book < ActiveRecord::Base
  enum status: [:proposed, :written, :published], _default: :published
end

Book.new.status # => "published"

Ryuta Kamizono

Deprecate YAML loading from legacy format older than Rails 5.0.
Ryuta Kamizono
Added the setting ActiveRecord::Base.immutable_strings_by_default, which
allows you to specify that all string columns should be frozen unless
otherwise specified. This will reduce memory pressure for applications which
do not generally mutate string properties of Active Record objects.
Sean Griffin, Ryuta Kamizono
Deprecate map! and collect! on ActiveRecord::Result.
Ryuta Kamizono

Support relation.and for intersection as Set theory.

david_and_mary = Author.where(id: [david, mary])
mary_and_bob   = Author.where(id: [mary, bob])

david_and_mary.merge(mary_and_bob) # => [mary, bob]

david_and_mary.and(mary_and_bob) # => [mary]
david_and_mary.or(mary_and_bob)  # => [david, mary, bob]

Ryuta Kamizono

Merging conditions on the same column no longer maintain both conditions,
and will be consistently replaced by the latter condition in Rails 6.2.
To migrate to Rails 6.2's behavior, use relation.merge(other, rewhere: true).

# Rails 6.1 (IN clause is replaced by merger side equality condition)
Author.where(id: [david.id, mary.id]).merge(Author.where(id: bob)) # => [bob]

# Rails 6.1 (both conflict conditions exists, deprecated)
Author.where(id: david.id..mary.id).merge(Author.where(id: bob)) # => []

# Rails 6.1 with rewhere to migrate to Rails 6.2's behavior
Author.where(id: david.id..mary.id).merge(Author.where(id: bob), rewhere: true) # => [bob]

# Rails 6.2 (same behavior with IN clause, mergee side condition is consistently replaced)
Author.where(id: [david.id, mary.id]).merge(Author.where(id: bob)) # => [bob]
Author.where(id: david.id..mary.id).merge(Author.where(id: bob)) # => [bob]

Ryuta Kamizono

Do not mark Postgresql MAC address and UUID attributes as changed when the assigned value only varies by case.
Peter Fry
Resolve issue with insert_all unique_by option when used with expression index.
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.
Usage:
```
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
```
Fixes #39516.
Austen Madden

Add basic support for CHECK constraints to database migrations.

Usage:

add_check_constraint :products, "price > 0", name: "price_check"
remove_check_constraint :products, name: "price_check"

fatkodima

Add ActiveRecord::Base.strict_loading_by_default and ActiveRecord::Base.strict_loading_by_default=
to enable/disable strict_loading mode by default for a model. The configuration's value is
inheritable by subclasses, but they can override that value and it will not impact parent class.
Usage:
```
class Developer < ApplicationRecord
  self.strict_loading_by_default = true

  has_many :projects
end

dev = Developer.first
dev.projects.first
# => ActiveRecord::StrictLoadingViolationError Exception: Developer is marked as strict_loading and Project cannot be lazily loaded.
```
bogdanvlviv
Deprecate passing an Active Record object to quote/type_cast directly.
Ryuta Kamizono

Default engine ENGINE=InnoDB is no longer dumped to make schema more agnostic.

Before:

create_table "accounts", options: "ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci", force: :cascade do |t|
end

After:

create_table "accounts", charset: "utf8mb4", collation: "utf8mb4_0900_ai_ci", force: :cascade do |t|
end

Ryuta Kamizono

Added delegated type as an alternative to single-table inheritance for representing class hierarchies.
See ActiveRecord::DelegatedType for the full description.
DHH

Deprecate aggregations with group by duplicated fields.

To migrate to Rails 6.2's behavior, use uniq!(:group) to deduplicate group fields.

accounts = Account.group(:firm_id)

# duplicated group fields, deprecated.
accounts.merge(accounts.where.not(credit_limit: nil)).sum(:credit_limit)
# => {
#   [1, 1] => 50,
#   [2, 2] => 60
# }

# use `uniq!(:group)` to deduplicate group fields.
accounts.merge(accounts.where.not(credit_limit: nil)).uniq!(:group).sum(:credit_limit)
# => {
#   1 => 50,
#   2 => 60
# }

Ryuta Kamizono

Deprecate duplicated query annotations.

To migrate to Rails 6.2's behavior, use uniq!(:annotate) to deduplicate query annotations.

accounts = Account.where(id: [1, 2]).annotate("david and mary")

# duplicated annotations, deprecated.
accounts.merge(accounts.rewhere(id: 3))
# SELECT accounts.* FROM accounts WHERE accounts.id = 3 /* david and mary */ /* david and mary */

# use `uniq!(:annotate)` to deduplicate annotations.
accounts.merge(accounts.rewhere(id: 3)).uniq!(:annotate)
# SELECT accounts.* FROM accounts WHERE accounts.id = 3 /* david and mary */

Ryuta Kamizono

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

Support merging option :rewhere to allow mergee side condition to be replaced exactly.

david_and_mary = Author.where(id: david.id..mary.id)

# both conflict conditions exists
david_and_mary.merge(Author.where(id: bob)) # => []

# mergee side condition is replaced by rewhere
david_and_mary.merge(Author.rewhere(id: bob)) # => [bob]

# mergee side condition is replaced by rewhere option
david_and_mary.merge(Author.where(id: bob), rewhere: true) # => [bob]

Ryuta Kamizono

Add support for finding records based on signed ids, which are tamper-proof, verified ids that can be
set to expire and scoped with a purpose. This is particularly useful for things like password reset
or email verification, where you want the bearer of the signed id to be able to interact with the
underlying record, but usually only within a certain time period.

signed_id = User.first.signed_id expires_in: 15.minutes, purpose: :password_reset

User.find_signed signed_id # => nil, since the purpose does not match

travel 16.minutes
User.find_signed signed_id, purpose: :password_reset # => nil, since the signed id has expired

travel_back
User.find_signed signed_id, purpose: :password_reset # => User.first

User.find_signed! "bad data" # => ActiveSupport::MessageVerifier::InvalidSignature

DHH

Support ALGORITHM = INSTANT DDL option for index operations on MySQL.
Ryuta Kamizono
Fix index creation to preserve index comment in bulk change table on MySQL.
Ryuta Kamizono

Allow unscope to be aware of table name qualified values.

It is possible to unscope only the column in the specified table.

posts = Post.joins(:comments).group(:"posts.hidden")
posts = posts.where("posts.hidden": false, "comments.hidden": false)

posts.count
# => { false => 10 }

# unscope both hidden columns
posts.unscope(where: :hidden).count
# => { false => 11, true => 1 }

# unscope only comments.hidden column
posts.unscope(where: :"comments.hidden").count
# => { false => 11 }

Ryuta Kamizono, Slava Korolev

Fix rewhere to truly overwrite collided where clause by new where clause.

steve = Person.find_by(name: "Steve")
david = Author.find_by(name: "David")

relation = Essay.where(writer: steve)

# Before
relation.rewhere(writer: david).to_a # => []

# After
relation.rewhere(writer: david).to_a # => [david]

Ryuta Kamizono

Inspect time attributes with subsec.

p Knot.create
=> #<Knot id: 1, created_at: "2016-05-05 01:29:47.116928000">

akinomaeni

Deprecate passing a column to type_cast.
Ryuta Kamizono
Deprecate in_clause_length and allowed_index_name_length in DatabaseLimits.
Ryuta Kamizono
Support bulk insert/upsert on relation to preserve scope values.
Josef Šimánek, Ryuta Kamizono
Preserve column comment value on changing column name on MySQL.
Islam Taha
Add support for if_exists option for removing an index.
The remove_index method can take an if_exists option. If this is set to true an error won't be raised if the index doesn't exist.
Eileen M. Uchitelle
Remove ibm_db, informix, mssql, oracle, and oracle12 Arel visitors which are not used in the code base.
Ryuta Kamizono
Prevent build_association from touching a parent record if the record isn't persisted for has_one associations.
Fixes #38219.
Josh Brody
Add support for if_not_exists option for adding index.
The add_index method respects if_not_exists option. If it is set to true
index won't be added.
Usage:
```
  add_index :users, :account_id, if_not_exists: true
```
The if_not_exists option passed to create_table also gets propagated to indexes
created within that migration so that if table and its indexes exist then there is no
attempt to create them again.
Prathamesh Sonpatki
Add ActiveRecord::Base#previously_new_record? to show if a record was new before the last save.
Tom Ward
Support descending order for find_each, find_in_batches, and in_batches.
Batch processing methods allow you to work with the records in batches, greatly reducing memory consumption, but records are always batched from oldest id to newest.
This change allows reversing the order, batching from newest to oldest. This is useful when you need to process newer batches of records first.
Pass order: :desc to yield batches in descending order. The default remains order: :asc.
```
Person.find_each(order: :desc) do |person|
  person.party_all_night!
end
```
Alexey Vasiliev
Fix insert_all with enum values.
Fixes #38716.
Joel Blum
Add support for db:rollback:name for multiple database applications.
Multiple database applications will now raise if db:rollback is call and recommend using the db:rollback:[NAME] to rollback migrations.
Eileen M. Uchitelle
Relation#pick now uses already loaded results instead of making another query.
Eugene Kenny
Deprecate using return, break or throw to exit a transaction block after writes.
Dylan Thacker-Smith
Dump the schema or structure of a database when calling db:migrate:name.
In previous versions of Rails, rails db:migrate would dump the schema of the database. In Rails 6, that holds true (rails db:migrate dumps all databases' schemas), but rails db:migrate:name does not share that behavior.
Going forward, calls to rails db:migrate:name will dump the schema (or structure) of the database being migrated.
Kyle Thompson
Reset the ActiveRecord::Base connection after rails db:migrate:name.
When rails db:migrate has finished, it ensures the ActiveRecord::Base connection is reset to its original configuration. Going forward, rails db:migrate:name will have the same behavior.
Kyle Thompson
Disallow calling connected_to on subclasses of ActiveRecord::Base.
Behavior has not changed here but the previous API could be misleading to people who thought it would switch connections for only that class. connected_to switches the context from which we are getting connections, not the connections themselves.
Eileen M. Uchitelle, John Crepezzi
Add support for horizontal sharding to connects_to and connected_to.
Applications can now connect to multiple shards and switch between their shards in an application. Note that the shard swapping is still a manual process as this change does not include an API for automatic shard swapping.
Usage:
Given the following configuration:
```
# config/database.yml
production:
  primary:
    database: my_database
  primary_shard_one:
    database: my_database_shard_one
```
Connect to multiple shards:
```
class ApplicationRecord < ActiveRecord::Base
  self.abstract_class = true

  connects_to shards: {
    default: { writing: :primary },
    shard_one: { writing: :primary_shard_one }
  }
```
Swap between shards in your controller / model code:
```
ActiveRecord::Base.connected_to(shard: :shard_one) do
  # Read from shard one
end
```
The horizontal sharding API also supports read replicas. See guides for more details.
Eileen M. Uchitelle, John Crepezzi

Deprecate spec_name in favor of name on database configurations.

The accessors for spec_name on configs_for and DatabaseConfig are deprecated. Please use name instead.

Deprecated behavior:

db_config = ActiveRecord::Base.configs_for(env_name: "development", spec_name: "primary")
db_config.spec_name

New behavior:

db_config = ActiveRecord::Base.configs_for(env_name: "development", name: "primary")
db_config.name

Eileen M. Uchitelle

Add additional database-specific rake tasks for multi-database users.

Previously, rails db:create, rails db:drop, and rails db:migrate were the only rails tasks that could operate on a single
database. For example:

rails db:create
rails db:create:primary
rails db:create:animals
rails db:drop
rails db:drop:primary
rails db:drop:animals
rails db:migrate
rails db:migrate:primary
rails db:migrate:animals

With these changes, rails db:schema:dump, rails db:schema:load, rails db:structure:dump, rails db:structure:load and
rails db:test:prepare can additionally operate on a single database. For example:

rails db:schema:dump
rails db:schema:dump:primary
rails db:schema:dump:animals
rails db:schema:load
rails db:schema:load:primary
rails db:schema:load:animals
rails db:structure:dump
rails db:structure:dump:primary
rails db:structure:dump:animals
rails db:structure:load
rails db:structure:load:primary
rails db:structure:load:animals
rails db:test:prepare
rails db:test:prepare:primary
rails db:test:prepare:animals

Kyle Thompson

Add support for strict_loading mode on association declarations.

Raise an error if attempting to load a record from an association that has been marked as strict_loading unless it was explicitly eager loaded.

Usage:

class Developer < ApplicationRecord
  has_many :projects, strict_loading: true
end

dev = Developer.first
dev.projects.first
# => ActiveRecord::StrictLoadingViolationError: The projects association is marked as strict_loading and cannot be lazily loaded.

Kevin Deisz

Add support for strict_loading mode to prevent lazy loading of records.
Raise an error if a parent record is marked as strict_loading and attempts to lazily load its associations. This is useful for finding places you may want to preload an association and avoid additional queries.
Usage:
```
dev = Developer.strict_loading.first
dev.audit_logs.to_a
# => ActiveRecord::StrictLoadingViolationError: Developer is marked as strict_loading and AuditLog cannot be lazily loaded.
```
Eileen M. Uchitelle, Aaron Patterson
Add support for PostgreSQL 11+ partitioned indexes when using upsert_all.
Sebastián Palma

Adds support for if_not_exists to add_column and if_exists to remove_column.

Applications can set their migrations to ignore exceptions raised when adding a column that already exists or when removing a column that does not exist.

Example Usage:

class AddColumnTitle < ActiveRecord::Migration[6.1]
  def change
    add_column :posts, :title, :string, if_not_exists: true
  end
end

class RemoveColumnTitle < ActiveRecord::Migration[6.1]
  def change
    remove_column :posts, :title, if_exists: true
  end
end

Eileen M. Uchitelle

Regexp-escape table name for MS SQL Server.
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.
Larry Reid
Store advisory locks on their own named connection.
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.
In order to fix this we are storing the advisory lock on a new connection with the connection specification name AdvisoryLockBase. The caveat is that we need to maintain at least 2 connections to a database while migrations are running in order to do this.
Eileen M. Uchitelle, John Crepezzi

Allow schema cache path to be defined in the database configuration file.

For example:

development:
  adapter: postgresql
  database: blog_development
  pool: 5
  schema_cache_path: tmp/schema/main.yml

Katrina Owen

Deprecate #remove_connection in favor of #remove_connection_pool when called on the handler.
#remove_connection is deprecated in order to support returning a DatabaseConfig object instead of a Hash. Use #remove_connection_pool, #remove_connection will be removed in 6.2.
Eileen M. Uchitelle, John Crepezzi
Deprecate #default_hash and it's alias #[] on database configurations.
Applications should use configs_for. #default_hash and #[] will be removed in 6.2.
Eileen M. Uchitelle, John Crepezzi
Add scale support to ActiveRecord::Validations::NumericalityValidator.
Gannon McGibbon
Find orphans by looking for missing relations through chaining where.missing:
Before:
```
Post.left_joins(:author).where(authors: { id: nil })
```
After:
```
Post.where.missing(:author)
```
Tom Rossi
Ensure :reading connections always raise if a write is attempted.
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.
Eileen M. Uchitelle
Deprecate "primary" as the connection_specification_name for ActiveRecord::Base.
"primary" has been deprecated as the connection_specification_name for ActiveRecord::Base in favor of using "ActiveRecord::Base". This change affects calls to ActiveRecord::Base.connection_handler.retrieve_connection and ActiveRecord::Base.connection_handler.remove_connection. If you're calling these methods with "primary", please switch to "ActiveRecord::Base".
Eileen M. Uchitelle, John Crepezzi
Add ActiveRecord::Validations::NumericalityValidator with
support for casting floats using a database columns' precision value.
Gannon McGibbon
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.
Aaron Lipman
Skip test database when running db:create or db:drop in development
with DATABASE_URL set.
Brian Buchalter

Don't allow mutations on the database configurations hash.

Freeze the configurations hash to disallow directly changing it. If applications need to change the hash, for example to create databases for parallelization, they should use the DatabaseConfig object directly.

Before:

@db_config = ActiveRecord::Base.configurations.configs_for(env_name: "test", spec_name: "primary")
@db_config.configuration_hash.merge!(idle_timeout: "0.02")

After:

@db_config = ActiveRecord::Base.configurations.configs_for(env_name: "test", spec_name: "primary")
config = @db_config.configuration_hash.merge(idle_timeout: "0.02")
db_config = ActiveRecord::DatabaseConfigurations::HashConfig.new(@db_config.env_name, @db_config.spec_name, config)

Eileen M. Uchitelle, John Crepezzi

Remove :connection_id from the sql.active_record notification.
Aaron Patterson, Rafael Mendonça França
The :name key will no longer be returned as part of DatabaseConfig#configuration_hash. Please use DatabaseConfig#owner_name instead.
Eileen M. Uchitelle, John Crepezzi
ActiveRecord's belongs_to_required_by_default flag can now be set per model.
You can now opt-out/opt-in specific models from having their associations required
by default.
This change is meant to ease the process of migrating all your models to have
their association required.
Edouard Chin
The connection_config method has been deprecated, please use connection_db_config instead which will return a DatabaseConfigurations::DatabaseConfig instead of a Hash.
Eileen M. Uchitelle, John Crepezzi
Retain explicit selections on the base model after applying includes and joins.
Resolves #34889.
Patrick Rebsch
The database kwarg is deprecated without replacement because it can't be used for sharding and creates an issue if it's used during a request. Applications that need to create new connections should use connects_to instead.
Eileen M. Uchitelle, John Crepezzi
Allow attributes to be fetched from Arel node groupings.
Jeff Emminger, Gannon McGibbon
A database URL can now contain a querystring value that contains an equal sign. This is needed to support passing PostgreSQL options.
Joshua Flanagan
Calling methods like establish_connection with a Hash which is invalid (eg: no adapter) will now raise an error the same way as connections defined in config/database.yml.
John Crepezzi
Specifying implicit_order_column now subsorts the records by primary key if available to ensure deterministic results.
Paweł Urbanek
where(attr => []) now loads an empty result without making a query.
John Hawthorn
Fixed the performance regression for primary_keys introduced MySQL 8.0.
Hiroyuki Ishii
Add support for belongs_to to has_many inversing.
Gannon McGibbon

Allow length configuration for has_secure_token method. The minimum length
is set at 24 characters.

Before:

has_secure_token :auth_token

After:

has_secure_token :default_token             # 24 characters
has_secure_token :auth_token, length: 36    # 36 characters
has_secure_token :invalid_token, length: 12 # => ActiveRecord::SecureToken::MinimumLengthError

Bernardo de Araujo

Deprecate DatabaseConfigurations#to_h. These connection hashes are still available via ActiveRecord::Base.configurations.configs_for.
Eileen Uchitelle, John Crepezzi
Add DatabaseConfig#configuration_hash to return database configuration hashes with symbol keys, and use all symbol-key configuration hashes internally. Deprecate DatabaseConfig#config which returns a String-keyed Hash with the same values.
John Crepezzi, Eileen Uchitelle

Allow column names to be passed to remove_index positionally along with other options.

Passing other options can be necessary to make remove_index correctly reversible.

Before:

add_index    :reports, :report_id               # => works
add_index    :reports, :report_id, unique: true # => works
remove_index :reports, :report_id               # => works
remove_index :reports, :report_id, unique: true # => ArgumentError

After:

remove_index :reports, :report_id, unique: true # => works

Eugene Kenny

Allow bulk ALTER statements to drop and recreate indexes with the same name.
Eugene Kenny
insert, insert_all, upsert, and upsert_all now clear the query cache.
Eugene Kenny
Call while_preventing_writes directly from connected_to.
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).
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.
Eileen M. Uchitelle
Improve detection of ActiveRecord::StatementTimeout with mysql2 adapter in the edge case when the query is terminated during filesort.
Kir Shatrov
Stop trying to read yaml file fixtures when loading Active Record fixtures.
Gannon McGibbon
Deprecate .reorder(nil) with .first / .first! taking non-deterministic result.
To continue taking non-deterministic result, use .take / .take! instead.
Ryuta Kamizono
Preserve user supplied joins order as much as possible.
Fixes #36761, #34328, #24281, #12953.
Ryuta Kamizono
Allow matches_regex and does_not_match_regexp on the MySQL Arel visitor.
James Pearson
Allow specifying fixtures to be ignored by setting ignore in YAML file's '_fixture' section.
Tongfei Gao
Make the DATABASE_URL env variable only affect the primary connection. Add new env variables for multiple databases.
John Crepezzi, Eileen Uchitelle
Add a warning for enum elements with 'not_' prefix.
```
class Foo
  enum status: [:sent, :not_sent]
end
```
Edu Depetris
Make currency symbols optional for money column type in PostgreSQL.
Joel Schneider
Add support for beginless ranges, introduced in Ruby 2.7.
Josh Goodall
Add database_exists? method to connection adapters to check if a database exists.
Guilherme Mansur
Loading the schema for a model that has no table_name raises a TableNotSpecified error.
Guilherme Mansur, Eugene Kenny
PostgreSQL: Fix GROUP BY with ORDER BY virtual count attribute.
Fixes #36022.
Ryuta Kamizono
Make ActiveRecord ConnectionPool.connections method thread-safe.
Fixes #36465.
Jeff Doering
Add support for multiple databases to rails db:abort_if_pending_migrations.
Mark Lee
Fix sqlite3 collation parsing when using decimal columns.
Martin R. Schuster
Fix invalid schema when primary key column has a comment.
Fixes #29966.
Guilherme Goettems Schneider
Fix table comment also being applied to the primary key column.
Guilherme Goettems Schneider
Allow generated create_table migrations to include or skip timestamps.
Michael Duchemin

Action View

Yield translated strings to calls of ActionView::FormBuilder#button
when a block is given.
Sean Doyle
Alias ActionView::Helpers::Tags::Label::LabelBuilder#translation to
#to_s so that form.label calls can yield that value to their blocks.
Sean Doyle
Rename the new TagHelper#class_names method to TagHelper#token_list,
and make the original available as an alias.
```
token_list("foo", "foo bar")
# => "foo bar"
```
Sean Doyle

ARIA Array and Hash attributes are treated as space separated DOMTokenList
values. This is useful when declaring lists of label text identifiers in
aria-labelledby or aria-describedby.

tag.input type: 'checkbox', name: 'published', aria: {
  invalid: @post.errors[:published].any?,
  labelledby: ['published_context', 'published_label'],
  describedby: { published_errors: @post.errors[:published].any? }
}
#=> <input
      type="checkbox" name="published" aria-invalid="true"
      aria-labelledby="published_context published_label"
      aria-describedby="published_errors"
    >

Sean Doyle

Remove deprecated escape_whitelist from ActionView::Template::Handlers::ERB.
Rafael Mendonça França
Remove deprecated find_all_anywhere from ActionView::Resolver.
Rafael Mendonça França
Remove deprecated formats from ActionView::Template::HTML.
Rafael Mendonça França
Remove deprecated formats from ActionView::Template::RawFile.
Rafael Mendonça França
Remove deprecated formats from ActionView::Template::Text.
Rafael Mendonça França
Remove deprecated find_file from ActionView::PathSet.
Rafael Mendonça França
Remove deprecated rendered_format from ActionView::LookupContext.
Rafael Mendonça França
Remove deprecated find_file from ActionView::ViewPaths.
Rafael Mendonça França
Require that ActionView::Base subclasses implement #compiled_method_container.
Rafael Mendonça França
Remove deprecated support to pass an object that is not a ActionView::LookupContext as the first argument
in ActionView::Base#initialize.
Rafael Mendonça França
Remove deprecated format argument ActionView::Base#initialize.
Rafael Mendonça França
Remove deprecated ActionView::Template#refresh.
Rafael Mendonça França
Remove deprecated ActionView::Template#original_encoding.
Rafael Mendonça França
Remove deprecated ActionView::Template#variants.
Rafael Mendonça França
Remove deprecated ActionView::Template#formats.
Rafael Mendonça França
Remove deprecated ActionView::Template#virtual_path=.
Rafael Mendonça França
Remove deprecated ActionView::Template#updated_at.
Rafael Mendonça França
Remove deprecated updated_at argument required on ActionView::Template#initialize.
Rafael Mendonça França
Make locals argument required on ActionView::Template#initialize.
Rafael Mendonça França
Remove deprecated ActionView::Template.finalize_compiled_template_methods.
Rafael Mendonça França
Remove deprecated config.action_view.finalize_compiled_template_methods
Rafael Mendonça França
Remove deprecated support to calling ActionView::ViewPaths#with_fallback with a block.
Rafael Mendonça França
Remove deprecated support to passing absolute paths to render template:.
Rafael Mendonça França
Remove deprecated support to passing relative paths to render file:.
Rafael Mendonça França
Remove support to template handlers that don't accept two arguments.
Rafael Mendonça França
Remove deprecated pattern argument in ActionView::Template::PathResolver.
Rafael Mendonça França
Remove deprecated support to call private methods from object in some view helpers.
Rafael Mendonça França
ActionView::Helpers::TranslationHelper#translate accepts a block, yielding
the translated text and the fully resolved translation key:
```
<%= translate(".relative_key") do |translation, resolved_key| %>
  <span title="<%= resolved_key %>"><%= translation %></span>
<% end %>
```
Sean Doyle
Ensure cache fragment digests include all relevant template dependencies when
fragments are contained in a block passed to the render helper. Remove the
virtual_path keyword arguments found in CacheHelper as they no longer possess
any function following 1581cab.
Fixes #38984.
Aaron Lipman
Deprecate config.action_view.raise_on_missing_translations in favor of
config.i18n.raise_on_missing_translations.
New generalized configuration option now determines whether an error should be raised
for missing translations in controllers and views.
fatkodima
Instrument layout rendering in TemplateRenderer#render_with_layout as render_layout.action_view,
and include (when necessary) the layout's virtual path in notification payloads for collection and partial renders.
Zach Kemp
ActionView::Base.annotate_rendered_view_with_filenames annotates HTML output with template file names.
Joel Hawksley, Aaron Patterson
ActionView::Helpers::TranslationHelper#translate returns nil when
passed default: nil without a translation matching I18n#translate.
Stefan Wrobel
OptimizedFileSystemResolver prefers template details in order of locale,
formats, variants, handlers.
Iago Pimenta
Added class_names helper to create a CSS class value with conditional classes.
Joel Hawksley, Aaron Patterson
Add support for conditional values to TagBuilder.
Joel Hawksley

ActionView::Helpers::FormOptionsHelper#select should mark option for nil as selected.

@post = Post.new
@post.category = nil

# Before
select("post", "category", none: nil, programming: 1, economics: 2)
# =>
# <select name="post[category]" id="post_category">
#   <option value="">none</option>
#  <option value="1">programming</option>
#  <option value="2">economics</option>
# </select>

# After
select("post", "category", none: nil, programming: 1, economics: 2)
# =>
# <select name="post[category]" id="post_category">
#   <option selected="selected" value="">none</option>
#  <option value="1">programming</option>
#  <option value="2">economics</option>
# </select>

bogdanvlviv

Log lines for partial renders and started template renders are now
emitted at the DEBUG level instead of INFO.
Completed template renders are still logged at the INFO level.
DHH
ActionView::Helpers::SanitizeHelper: support rails-html-sanitizer 1.1.0.
Juanito Fatas
Added phone_to helper method to create a link from mobile numbers.
Pietro Moro
annotated_source_code returns an empty array so TemplateErrors without a
template in the backtrace are surfaced properly by DebugExceptions.
Guilherme Mansur, Kasper Timm Hansen
Add autoload for SyntaxErrorInTemplate so syntax errors are correctly raised by DebugExceptions.
Guilherme Mansur, Gannon McGibbon
RenderingHelper supports rendering objects that respond_to? :render_in.
Joel Hawksley, Natasha Umer, Aaron Patterson, Shawn Allen, Emily Plummer, Diana Mounter, John Hawthorn, Nathan Herald, Zaid Zawaideh, Zach Ahn
Fix select_tag so that it doesn't change options when include_blank is present.
Younes SERRAJ

Action Pack

Allow ActionDispatch::HostAuthorization to exclude specific requests.
Host Authorization checks can be skipped for specific requests. This allows for health check requests to be permitted for requests with missing or non-matching host headers.
Chris Bisnett
Add config.action_dispatch.request_id_header to allow changing the name of
the unique X-Request-Id header
Arlston Fernandes
Deprecate config.action_dispatch.return_only_media_type_on_content_type.
Rafael Mendonça França
Change ActionDispatch::Response#content_type to return the full Content-Type header.
Rafael Mendonça França
Remove deprecated ActionDispatch::Http::ParameterFilter.
Rafael Mendonça França
Added support for exclusive no-store Cache-Control header.
If no-store is set on Cache-Control header it is exclusive (all other cache directives are dropped).
Chris Kruger
Catch invalid UTF-8 parameters for POST requests and respond with BadRequest.
Additionally, perform #set_binary_encoding in ActionDispatch::Http::Request#GET and
ActionDispatch::Http::Request#POST prior to validating encoding.
Adrianna Chang
Allow assert_recognizes routing assertions to work on mounted root routes.
Gannon McGibbon
Change default redirection status code for non-GET/HEAD requests to 308 Permanent Redirect for ActionDispatch::SSL.
Alan Tan, Oz Ben-David
Fix follow_redirect! to follow redirection with same HTTP verb when following
a 308 redirection.
Alan Tan
When multiple domains are specified for a cookie, a domain will now be
chosen only if it is equal to or is a superdomain of the request host.
Jonathan Hefner
ActionDispatch::Static handles precompiled Brotli (.br) files.
Adds to existing support for precompiled gzip (.gz) files.
Brotli files are preferred due to much better compression.
When the browser requests /some.js with Accept-Encoding: br,
we check for public/some.js.br and serve that file, if present, with
Content-Encoding: br and Vary: Accept-Encoding headers.
Ryan Edward Hall, Jeremy Daer
Add raise_on_missing_translations support for controllers.
This configuration determines whether an error should be raised for missing translations.
It can be enabled through config.i18n.raise_on_missing_translations. Note that described
configuration also affects raising error for missing translations in views.
fatkodima
Added compact and compact! to ActionController::Parameters.
Eugene Kenny
Calling each_pair or each_value on an ActionController::Parameters
without passing a block now returns an enumerator.
Eugene Kenny
fixture_file_upload now uses path relative to file_fixture_path
Previously the path had to be relative to fixture_path.
You can change your existing code as follow:
```
# Before
fixture_file_upload('files/dog.png')

# After
fixture_file_upload('dog.png')
```
Edouard Chin
Remove deprecated force_ssl at the controller level.
Rafael Mendonça França
The +helper+ class method for controllers loads helper modules specified as
strings/symbols with String#constantize instead of require_dependency.
Remember that support for strings/symbols is only a convenient API. You can
always pass a module object:
```
helper UtilsHelper
```
which is recommended because it is simple and direct. When a string/symbol
is received, helper just manipulates and inflects the argument to obtain
that same module object.
Xavier Noria, Jean Boussier
Correctly identify the entire localhost IPv4 range as trusted proxy.
Nick Soracco
url_for will now use "https://" as the default protocol when
Rails.application.config.force_ssl is set to true.
Jonathan Hefner
Accept and default to base64_urlsafe CSRF tokens.
Base64 strict-encoded CSRF tokens are not inherently websafe, which makes
them difficult to deal with. For example, the common practice of sending
the CSRF token to a browser in a client-readable cookie does not work properly
out of the box: the value has to be url-encoded and decoded to survive transport.
Now, we generate Base64 urlsafe-encoded CSRF tokens, which are inherently safe
to transport. Validation accepts both urlsafe tokens, and strict-encoded tokens
for backwards compatibility.
Scott Blum
Support rolling deploys for cookie serialization/encryption changes.
In a distributed configuration like rolling update, users may observe
both old and new instances during deployment. Users may be served by a
new instance and then by an old instance.
That means when the server changes cookies_serializer from :marshal
to :hybrid or the server changes use_authenticated_cookie_encryption
from false to true, users may lose their sessions if they access the
server during deployment.
We added fallbacks to downgrade the cookie format when necessary during
deployment, ensuring compatibility on both old and new instances.
Masaki Hara
ActionDispatch::Request.remote_ip has ip address even when all sites are trusted.
Before, if all X-Forwarded-For sites were trusted, the remote_ip would default to 127.0.0.1.
Now, the furthest proxy site is used. e.g.: It now gives an ip address when using curl from the load balancer.
Keenan Brock
Fix possible information leak / session hijacking vulnerability.
The ActionDispatch::Session::MemcacheStore is still vulnerable given it requires the
gem dalli to be updated as well.
CVE-2019-16782.
Include child session assertion count in ActionDispatch::IntegrationTest.
IntegrationTest#open_session uses dup to create the new session, which
meant it had its own copy of @assertions. This prevented the assertions
from being correctly counted and reported.
Child sessions now have their attr_accessor overridden to delegate to the
root session.
Fixes #32142.
Sam Bostock
Add SameSite protection to every written cookie.
Enabling SameSite cookie protection is an addition to CSRF protection,
where cookies won't be sent by browsers in cross-site POST requests when set to :lax.
:strict disables cookies being sent in cross-site GET or POST requests.
Passing :none disables this protection and is the same as previous versions albeit a ; SameSite=None is appended to the cookie.
See upgrade instructions in config/initializers/new_framework_defaults_6_1.rb.
More info here
NB: Technically already possible as Rack supports SameSite protection, this is to ensure it's applied to all cookies
Cédric Fabianski
Bring back the feature that allows loading external route files from the router.
This feature existed back in 2012 but got reverted with the incentive that
https://github.com/rails/routing_concerns was a better approach. Turned out
that this wasn't fully the case and loading external route files from the router
can be helpful for applications with a really large set of routes.
Without this feature, application needs to implement routes reloading
themselves and it's not straightforward.
```
# config/routes.rb

Rails.application.routes.draw do
  draw(:admin)
end

# config/routes/admin.rb

get :foo, to: 'foo#bar'
```
Yehuda Katz, Edouard Chin
Fix system test driver option initialization for non-headless browsers.
glaszig
redirect_to.action_controller notifications now include the ActionDispatch::Request in
their payloads as :request.
Austin Story
respond_to#any no longer returns a response's Content-Type based on the
request format but based on the block given.
Example:
```
  def my_action
    respond_to do |format|
      format.any { render(json: { foo: 'bar' }) }
    end
  end

  get('my_action.csv')
```
The previous behaviour was to respond with a text/csv Content-Type which
is inaccurate since a JSON response is being rendered.
Now it correctly returns a application/json Content-Type.
Edouard Chin
Replaces (back)slashes in failure screenshot image paths with dashes.
If a failed test case contained a slash or a backslash, a screenshot would be created in a
nested directory, causing issues with tmp:clear.
Damir Zekic
Add params.member? to mimic Hash behavior.
Younes Serraj
process_action.action_controller notifications now include the following in their payloads:
- :request - the ActionDispatch::Request
- :response - the ActionDispatch::Response
George Claghorn
Updated ActionDispatch::Request.remote_ip setter to clear set the instance
remote_ip to nil before setting the header that the value is derived
from.
Fixes #37383.
Norm Provost

ActionController::Base.log_at allows setting a different log level per request.

# Use the debug level if a particular cookie is set.
class ApplicationController < ActionController::Base
  log_at :debug, if: -> { cookies[:debug] }
end

George Claghorn

Allow system test screen shots to be taken more than once in
a test by prefixing the file name with an incrementing counter.
Add an environment variable RAILS_SYSTEM_TESTING_SCREENSHOT_HTML to
enable saving of HTML during a screenshot in addition to the image.
This uses the same image name, with the extension replaced with .html
Tom Fakes
Add Vary: Accept header when using Accept header for response.
For some requests like /users/1, Rails uses requests' Accept
header to determine what to return. And if we don't add Vary
in the response header, browsers might accidentally cache different
types of content, which would cause issues: e.g. javascript got displayed
instead of html content. This PR fixes these issues by adding Vary: Accept
in these types of requests. For more detailed problem description, please read:
#36213
Fixes #25842.
Stan Lo
Fix IntegrationTest follow_redirect! to follow redirection using the same HTTP verb when following
a 307 redirection.
Edouard Chin
System tests require Capybara 3.26 or newer.
George Claghorn
Reduced log noise handling ActionController::RoutingErrors.
Alberto Fernández-Capel

Add DSL for configuring HTTP Feature Policy.

This new DSL provides a way to configure an HTTP Feature Policy at a
global or per-controller level. Full details of HTTP Feature Policy
specification and guidelines can be found at MDN:

https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Feature-Policy

Example global policy:

Rails.application.config.feature_policy do |f|
  f.camera      :none
  f.gyroscope   :none
  f.microphone  :none
  f.usb         :none
  f.fullscreen  :self
  f.payment     :self, "https://secure.example.com"
end

Example controller level policy:

class PagesController < ApplicationController
  feature_policy do |p|
    p.geolocation "https://example.com"
  end
end

Jacob Bednarz

Add the ability to set the CSP nonce only to the specified directives.
Fixes #35137.
Yuji Yaginuma
Keep part when scope option has value.
When a route was defined within an optional scope, if that route didn't
take parameters the scope was lost when using path helpers. This commit
ensures scope is kept both when the route takes parameters or when it
doesn't.
Fixes #33219.
Alberto Almagro
Added deep_transform_keys and deep_transform_keys! methods to ActionController::Parameters.
Gustavo Gutierrez
Calling ActionController::Parameters#transform_keys/! without a block now returns
an enumerator for the parameters instead of the underlying hash.
Eugene Kenny
Fix strong parameters blocks all attributes even when only some keys are invalid (non-numerical).
It should only block invalid key's values instead.
Stan Lo

Active Job

Recover nano precision when serializing Time, TimeWithZone and DateTime objects.
Alan Tan
Deprecate config.active_job.return_false_on_aborted_enqueue.
Rafael Mendonça França
Return false when enqueuing a job is aborted.
Rafael Mendonça França
While using perform_enqueued_jobs test helper enqueued jobs must be stored for the later check with
assert_enqueued_with.
Dmitry Polushkin
ActiveJob::TestCase#perform_enqueued_jobs without a block removes performed jobs from the queue.
That way the helper can be called multiple times and not perform a job invocation multiple times.
```
def test_jobs
  HelloJob.perform_later("rafael")
  perform_enqueued_jobs # only runs with "rafael"
  HelloJob.perform_later("david")
  perform_enqueued_jobs # only runs with "david"
end
```
Étienne Barrié
ActiveJob::TestCase#perform_enqueued_jobs will no longer perform retries:
When calling perform_enqueued_jobs without a block, the adapter will
now perform jobs that are already in the queue. Jobs that will end up in
the queue afterwards won't be performed.
This change only affects perform_enqueued_jobs when no block is given.
Edouard Chin
Add queue name support to Que adapter.
Brad Nauta, Wojciech Wnętrzak
Don't run after_enqueue and after_perform callbacks if the callback chain is halted.
```
class MyJob < ApplicationJob
  before_enqueue { throw(:abort) }
  after_enqueue { # won't enter here anymore }
end
```
after_enqueue and after_perform callbacks will no longer run if the callback chain is halted.
This behaviour is a breaking change and won't take effect until Rails 6.2.
To enable this behaviour in your app right now, you can add in your app's configuration file
config.active_job.skip_after_callbacks_if_terminated = true.
Edouard Chin
Fix enqueuing and performing incorrect logging message.
Jobs will no longer always log "Enqueued MyJob" or "Performed MyJob" when they actually didn't get enqueued/performed.
```
  class MyJob < ApplicationJob
    before_enqueue { throw(:abort) }
  end

  MyJob.perform_later # Will no longer log "Enqueued MyJob" since job wasn't even enqueued through adapter.
```
A new message will be logged in case a job couldn't be enqueued, either because the callback chain was halted or
because an exception happened during enqueing. (i.e. Redis is down when you try to enqueue your job)
Edouard Chin
Add an option to disable logging of the job arguments when enqueuing and executing the job.
```
class SensitiveJob < ApplicationJob
  self.log_arguments = false

  def perform(my_sensitive_argument)
  end
end
```
When dealing with sensitive arguments as password and tokens it is now possible to configure the job
to not put the sensitive argument in the logs.
Rafael Mendonça França
Changes in queue_name_prefix of a job no longer affects all other jobs.
Fixes #37084.
Lucas Mansur
Allow Class and Module instances to be serialized.
Kevin Deisz
Log potential matches in assert_enqueued_with and assert_performed_with.
Gareth du Plooy
Add at argument to the perform_enqueued_jobs test helper.
John Crepezzi, Eileen Uchitelle
assert_enqueued_with and assert_performed_with can now test jobs with relative delay.
Vlado Cingel
Add jitter to ActiveJob::Exceptions.retry_on.
ActiveJob::Exceptions.retry_on now uses a random amount of jitter in order to
prevent the thundering herd effect. Defaults to
15% (represented as 0.15) but overridable via the :jitter option when using retry_on.
Jitter is applied when an Integer, ActiveSupport::Duration or :exponentially_longer, is passed to the wait argument in retry_on.
```
retry_on(MyError, wait: :exponentially_longer, jitter: 0.30)
```
Anthony Ross

Action Mailer

Remove deprecated ActionMailer::Base.receive in favor of Action Mailbox.
Rafael Mendonça França
Fix ActionMailer assertions don't work for parameterized mail with legacy delivery job.
bogdanvlviv
Added email_address_with_name to properly escape addresses with names.
Sunny Ripert

Action Cable

ActionCable::Connection::Base now allows intercepting unhandled exceptions
with rescue_from before they are logged, which is useful for error reporting
tools and other integrations.
Justin Talbott
Add ActionCable::Channel#stream_or_reject_for to stream if record is present, otherwise reject the connection
Atul Bhosale
Add ActionCable::Channel#stop_stream_from and #stop_stream_for to unsubscribe from a specific stream.
Zhang Kang
Add PostgreSQL subscription connection identificator.
Now you can distinguish Action Cable PostgreSQL subscription connections among others.
Also, you can set custom id in cable.yml configuration.
```
SELECT application_name FROM pg_stat_activity;
/*
    application_name
------------------------
psql
ActionCable-PID-42
(2 rows)
*/
```
Sergey Ponomarev
Subscription confirmations and rejections are now logged at the DEBUG level instead of INFO.
DHH

Active Storage

Remove deprecated support to pass :combine_options operations to ActiveStorage::Transformers::ImageProcessing.
Rafael Mendonça França
Remove deprecated ActiveStorage::Transformers::MiniMagickTransformer.
Rafael Mendonça França
Remove deprecated config.active_storage.queue.
Rafael Mendonça França
Remove deprecated ActiveStorage::Downloading.
Rafael Mendonça França
Add per-environment configuration support
Pietro Moro
The Poppler PDF previewer renders a preview image using the original
document's crop box rather than its media box, hiding print margins. This
matches the behavior of the MuPDF previewer.
Vincent Robert
Touch parent model when an attachment is purged.
Víctor Pérez Rodríguez
Files can now be served by proxying them from the underlying storage service
instead of redirecting to a signed service URL. Use the
rails_storage_proxy_path and _url helpers to proxy an attached file:
```
<%= image_tag rails_storage_proxy_path(@user.avatar) %>
```
To proxy by default, set config.active_storage.resolve_model_to_route:
```
# Proxy attached files instead.
config.active_storage.resolve_model_to_route = :rails_storage_proxy
```
```
<%= image_tag @user.avatar %>
```
To redirect to a signed service URL when the default file serving strategy
is set to proxying, use the rails_storage_redirect_path and _url helpers:
```
<%= image_tag rails_storage_redirect_path(@user.avatar) %>
```
Jonathan Fleckenstein
Add config.active_storage.web_image_content_types to allow applications
to add content types (like image/webp) in which variants can be processed,
instead of letting those images be converted to the fallback PNG format.
Jeroen van Haperen
Add support for creating variants of WebP images out of the box.
Dino Maric
Only enqueue analysis jobs for blobs with non-null analyzer classes.
Gannon McGibbon
Previews are created on the same service as the original blob.
Peter Zhu
Remove unused disposition and content_type query parameters for DiskService.
Peter Zhu
Use DiskController for both public and private files.
DiskController is able to handle multiple services by adding a
service_name field in the generated URL in DiskService.
Peter Zhu
Variants are tracked in the database to avoid existence checks in the storage service.
George Claghorn
Deprecate service_url methods in favour of url.
Deprecate Variant#service_url and Preview#service_url to instead use
#url method to be consistent with Blob.
Peter Zhu
Permanent URLs for public storage blobs.
Services can be configured in config/storage.yml with a new key
public: true | false to indicate whether a service holds public
blobs or private blobs. Public services will always return a permanent URL.
Deprecates Blob#service_url in favor of Blob#url.
Peter Zhu
Make services aware of configuration names.
Gannon McGibbon
The Content-Type header is set on image variants when they're uploaded to third-party storage services.
Kyle Ribordy

Allow storage services to be configured per attachment.

class User < ActiveRecord::Base
  has_one_attached :avatar, service: :s3
end

class Gallery < ActiveRecord::Base
  has_many_attached :photos, service: :s3
end

Dmitry Tsepelev

You can optionally provide a custom blob key when attaching a new file:
```
user.avatar.attach key: "avatars/#{user.id}.jpg",
  io: io, content_type: "image/jpeg", filename: "avatar.jpg"
```
Active Storage will store the blob's data on the configured service at the provided key.
George Claghorn
Replace Blob.create_after_upload! with Blob.create_and_upload! and deprecate the former.
create_after_upload! has been removed since it could lead to data
corruption by uploading to a key on the storage service which happened to
be already taken. Creating the record would then correctly raise a
database uniqueness exception but the stored object would already have
overwritten another. create_and_upload! swaps the order of operations
so that the key gets reserved up-front or the uniqueness error gets raised,
before the upload to a key takes place.
Julik Tarkhanov
Set content disposition in direct upload using filename and disposition parameters to ActiveStorage::Service#headers_for_direct_upload.
Peter Zhu
Allow record to be optionally passed to blob finders to make sharding
easier.
Gannon McGibbon
Switch from azure-storage gem to azure-storage-blob gem for Azure service.
Peter Zhu
Add config.active_storage.draw_routes to disable Active Storage routes.
Gannon McGibbon
Image analysis is skipped if ImageMagick returns an error.
ActiveStorage::Analyzer::ImageAnalyzer#metadata would previously raise a
MiniMagick::Error, which caused persistent ActiveStorage::AnalyzeJob
failures. It now logs the error and returns {}, resulting in no metadata
being added to the offending image blob.
George Claghorn
Method calls on singular attachments return nil when no file is attached.
Previously, assuming the following User model, user.avatar.filename would
raise a Module::DelegationError if no avatar was attached:
```
class User < ApplicationRecord
  has_one_attached :avatar
end
```
They now return nil.
Matthew Tanous
The mirror service supports direct uploads.
New files are directly uploaded to the primary service. When a
directly-uploaded file is attached to a record, a background job is enqueued
to copy it to each secondary service.
Configure the queue used to process mirroring jobs by setting
config.active_storage.queues.mirror. The default is :active_storage_mirror.
George Claghorn
The S3 service now permits uploading files larger than 5 gigabytes.
When uploading a file greater than 100 megabytes in size, the service
transparently switches to multipart uploads
using a part size computed from the file's total size and S3's part count limit.
No application changes are necessary to take advantage of this feature. You
can customize the default 100 MB multipart upload threshold in your S3
service's configuration:
```
production:
  service: s3
  access_key_id: <%= Rails.application.credentials.dig(:aws, :access_key_id) %>
  secret_access_key: <%= Rails.application.credentials.dig(:aws, :secret_access_key) %>
  region: us-east-1
  bucket: my-bucket
  upload:
    multipart_threshold: <%= 250.megabytes %>
```
George Claghorn

Action Mailbox

Sendgrid ingress now passes through the envelope recipient as X-Original-To.
Mark Haussmann
Update Mandrill inbound email route to respond appropriately to HEAD requests for URL health checks from Mandrill.
Bill Cromie
Add way to deliver emails via source instead of filling out a form through the conductor interface.
DHH
Mailgun ingress now passes through the envelope recipient as X-Original-To.
Rikki Pitt
Deprecate Rails.application.credentials.action_mailbox.api_key and MAILGUN_INGRESS_API_KEY in favor of Rails.application.credentials.action_mailbox.signing_key and MAILGUN_INGRESS_SIGNING_KEY.
Matthijs Vos
Allow easier creation of multi-part emails from the create_inbound_email_from_mail and receive_inbound_email_from_mail test helpers.
Michael Herold
Fix Bcc header not being included with emails from create_inbound_email_from test helpers.
jduff
Add ApplicationMailbox.mailbox_for to expose mailbox routing.
James Dabbs

Action Text

Declare ActionText::FixtureSet.attachment to generate an
<action-text-attachment sgid="..."></action-text-attachment> element with
a valid sgid attribute.

hello_world_review_content:
  record: hello_world (Review)
  name: content
  body: <p><%= ActionText::FixtureSet.attachment("messages", :hello_world) %> is great!</p>

Sean Doyle

Locate fill_in_rich_text_area by <label> text
In addition to searching for <trix-editor> elements with the appropriate
aria-label attribute, also support locating elements that match the
corresponding <label> element's text.
Sean Doyle

Be able to add a default value to rich_text_area.

form.rich_text_area :content, value: "<h1>Hello world</h1>"
#=> <input type="hidden" name="message[content]" id="message_content_trix_input_message_1" value="<h1>Hello world</h1>">

Paulo Ancheta

Add method to confirm rich text content existence by adding ? after rich
text attribute.
```
message = Message.create!(body: "<h1>Funny times!</h1>")
message.body? #=> true
```
Kyohei Toyoda

The fill_in_rich_text_area system test helper locates a Trix editor
and fills it in with the given HTML.

# <trix-editor id="message_content" ...></trix-editor>
fill_in_rich_text_area "message_content", with: "Hello <em>world!</em>"

# <trix-editor placeholder="Your message here" ...></trix-editor>
fill_in_rich_text_area "Your message here", with: "Hello <em>world!</em>"

# <trix-editor aria-label="Message content" ...></trix-editor>
fill_in_rich_text_area "Message content", with: "Hello <em>world!</em>"

# <input id="trix_input_1" name="message[content]" type="hidden">
# <trix-editor input="trix_input_1"></trix-editor>
fill_in_rich_text_area "message[content]", with: "Hello <em>world!</em>"

George Claghorn

Railties

Added Railtie#server hook called when Rails starts a server.
This is useful in case your application or a library needs to run
another process next to the Rails server. This is quite common in development
for instance to run the Webpack or the React server.
It can be used like this:
```
  class MyRailtie < Rails::Railtie
    server do
      WebpackServer.run
    end
  end
```
Edouard Chin
Remove deprecated rake dev:cache tasks.
Rafael Mendonça França
Remove deprecated rake routes tasks.
Rafael Mendonça França
Remove deprecated rake initializers tasks.
Rafael Mendonça França
Remove deprecated support for using the HOST environment variable to specify the server IP.
Rafael Mendonça França
Remove deprecated server argument from the rails server command.
Rafael Mendonça França
Remove deprecated SOURCE_ANNOTATION_DIRECTORIES environment variable support from rails notes.
Rafael Mendonça França
Remove deprecated connection option in the rails dbconsole command.
Rafael Mendonça França
Remove depreated rake notes tasks.
Rafael Mendonça França
Return a 405 Method Not Allowed response when a request uses an unknown HTTP method.
Fixes #38998.
Loren Norman
Make railsrc file location xdg-specification compliant
rails new will now look for the default railsrc file at
$XDG_CONFIG_HOME/rails/railsrc (or ~/.config/rails/railsrc if
XDG_CONFIG_HOME is not set). If this file does not exist, rails new
will fall back to ~/.railsrc.
The fallback behaviour means this does not cause any breaking changes.
Nick Wolf
Change the default logging level from :debug to :info to avoid inadvertent exposure of personally
identifiable information (PII) in production environments.
Eric M. Payne
Automatically generate abstract class when using multiple databases.
When generating a scaffold for a multiple database application, Rails will now automatically generate the abstract class for the database when the database argument is passed. This abstract class will include the connection information for the writing configuration and any models generated for that database will automatically inherit from the abstract class.
Usage:
```
$ bin/rails generate scaffold Pet name:string --database=animals
```
Will create an abstract class for the animals connection.
```
class AnimalsRecord < ApplicationRecord
  self.abstract_class = true

  connects_to database: { writing: :animals }
end
```
And generate a Pet model that inherits from the new AnimalsRecord:
```
class Pet < AnimalsRecord
end
```
If you already have an abstract class and it follows a different pattern than Rails defaults, you can pass a parent class with the database argument.
```
$ bin/rails generate scaffold Pet name:string --database=animals --parent=SecondaryBase
```
This will ensure the model inherits from the SecondaryBase parent instead of AnimalsRecord
```
class Pet < SecondaryBase
end
```
Eileen M. Uchitelle, John Crepezzi
Accept params from url to prepopulate the Inbound Emails form in Rails conductor.
Chris Oliver
Create a new rails app using a minimal stack.
rails new cool_app --minimal
All the following are excluded from your minimal stack:
- action_cable
- action_mailbox
- action_mailer
- action_text
- active_job
- active_storage
- bootsnap
- jbuilder
- spring
- system_tests
- turbolinks
- webpack
Haroon Ahmed, DHH
Add default ENV variable option with BACKTRACE to turn off backtrace cleaning when debugging framework code in the
generated config/initializers/backtrace_silencers.rb.
BACKTRACE=1 ./bin/rails runner "MyClass.perform"
DHH
The autoloading guide for Zeitwerk mode documents how to autoload classes
during application boot in a safe way.
Haroon Ahmed, Xavier Noria
The classic autoloader starts its deprecation cycle.
New Rails projects are strongly discouraged from using classic, and we recommend that existing projects running on classic switch to zeitwerk mode when upgrading. Please check the Upgrading Ruby on Rails guide for tips.
Xavier Noria
Adds rails test:all for running all tests in the test directory.
This runs all test files in the test directory, including system tests.
Niklas Häusele
Add config.generators.after_generate for processing to generated files.
Register a callback that will get called right after generators has finished.
Yuji Yaginuma
Make test file patterns configurable via Environment variables
This makes test file patterns configurable via two environment variables:
DEFAULT_TEST, to configure files to test, and DEFAULT_TEST_EXCLUDE,
to configure files to exclude from testing.
These values were hardcoded before, which made it difficult to add
new categories of tests that should not be executed by default (e.g:
smoke tests).
Jorge Manrubia
No longer include rake rdoc task when generating plugins.
To generate docs, use the rdoc lib command instead.
Jonathan Hefner
Allow relative paths with trailing slashes to be passed to rails test.
Eugene Kenny
Add rack-mini-profiler gem to the default Gemfile.
rack-mini-profiler displays performance information such as SQL time and flame graphs.
It's enabled by default in development environment, but can be enabled in production as well.
See the gem README for information on how to enable it in production.
Osama Sayegh
rails stats will now count TypeScript files toward JavaScript stats.
Joshua Cody
Run git init when generating plugins.
Opt out with --skip-git.
OKURA Masafumi
Add benchmark generator.
Introduce benchmark generator to benchmark Rails applications.
rails generate benchmark opt_compare
This creates a benchmark file that uses benchmark-ips.
By default, two code blocks can be benchmarked using the before and after reports.
You can run the generated benchmark file using:
ruby script/benchmarks/opt_compare.rb
Kevin Jalbert, Gannon McGibbon
Cache compiled view templates when running tests by default.
When generating a new app without --skip-spring, caching classes is
disabled in environments/test.rb. This implicitly disables caching
view templates too. This change will enable view template caching by
adding this to the generated environments/test.rb:
```
config.action_view.cache_template_loading = true
```
Jorge Manrubia
Introduce middleware move operations.
With this change, you no longer need to delete and reinsert a middleware to
move it from one place to another in the stack:
```
config.middleware.move_before ActionDispatch::Flash, Magical::Unicorns
```
This will move the Magical::Unicorns middleware before
ActionDispatch::Flash. You can also move it after with:
```
config.middleware.move_after ActionDispatch::Flash, Magical::Unicorns
```
Genadi Samokovarov
Generators that inherit from NamedBase respect --force option.
Josh Brody
Allow configuration of eager_load behaviour for rake environment:
```
config.rake_eager_load
```
Defaults to false as per previous behaviour.
Thierry Joyal
Ensure Rails migration generator respects system-wide primary key config.
When rails is configured to use a specific primary key type:
```
config.generators do |g|
  g.orm :active_record, primary_key_type: :uuid
end
```
Previously:
```
$ bin/rails g migration add_location_to_users location:references
```
The references line in the migration would not have type: :uuid.
This change causes the type to be applied appropriately.
Louis-Michel Couture, Dermot Haughey
Deprecate Rails::DBConsole#config.
Rails::DBConsole#config is deprecated without replacement. Use Rails::DBConsole.db_config.configuration_hash instead.
Eileen M. Uchitelle, John Crepezzi

Rails.application.config_for merges shared configuration deeply.

# config/example.yml
shared:
  foo:
    bar:
      baz: 1
development:
  foo:
    bar:
      qux: 2

# Previously
Rails.application.config_for(:example)[:foo][:bar] #=> { qux: 2 }

# Now
Rails.application.config_for(:example)[:foo][:bar] #=> { baz: 1, qux: 2 }

Yuhei Kiriyama

Remove access to values in nested hashes returned by Rails.application.config_for via String keys.
```
# config/example.yml
development:
  options:
    key: value
```
```
Rails.application.config_for(:example).options
```
This used to return a Hash on which you could access values with String keys. This was deprecated in 6.0, and now doesn't work anymore.
Étienne Barrié
Configuration files for environments (config/environments/*.rb) are
now able to modify autoload_paths, autoload_once_paths, and
eager_load_paths.
As a consequence, applications cannot autoload within those files. Before, they technically could, but changes in autoloaded classes or modules had no effect anyway in the configuration because reloading does not reboot.
Ways to use application code in these files:
- Define early in the boot process a class that is not reloadable, from which the application takes configuration values that get passed to the framework.
```
# In config/application.rb, for example.
require "#{Rails.root}/lib/my_app/config"

# In config/environments/development.rb, for example.
config.foo = MyApp::Config.foo
```
- If the class has to be reloadable, then wrap the configuration code in a to_prepare block:
```
config.to_prepare do
  config.foo = MyModel.foo
end
```
  That assigns the latest MyModel.foo to config.foo when the application boots, and each time there is a reload. But whether that has an effect or not depends on the configuration point, since it is not uncommon for engines to read the application configuration during initialization and set their own state from them. That process happens only on boot, not on reloads, and if that is how config.foo worked, resetting it would have no effect in the state of the engine.
Allen Hsu & Xavier Noria
Support using environment variable to set pidfile.
Ben Thorner