sequel 4.26.0 → 5.37.0

data/doc/release_notes/5.5.0.txt

@@ -0,0 +1,61 @@
+= New Features
+
+* The defaults_setter plugin now supports a :cache option, which
+  will cache default values in the model object's values hash:
+
+    Model.plugin :defaults_setter
+    o = Model.new
+    o.column # => 1 # default value
+    o.values # => {}
+
+    Model.plugin :defaults_setter, cache: true
+    o = Model.new
+    o.column # => 1 # default value
+    o.values # => {:column => 1}
+
+* The pg_array extension now sets a :callable_default schema entry
+  for recognized empty array defaults.
+
+* The pg_hstore extension now sets a :callable_default schema entry
+  for recognized empty hstore defaults.
+
+* The pg_json extension now sets a :callable_default schema entry for
+  recognized empty json/jsonb array/hash defaults. 
+
+* The pg_inet extension now sets a :ruby_default schema entry for
+  recognized inet/cidr defaults.
+
+* The pg_range extension now sets a :ruby_default schema entry for
+  recognized range defaults.
+
+* The defaults_setter plugin will now give preference to a
+  :callable_default schema entry over a :ruby_default schema entry.
+  Combined with the other changes listed above, this makes default
+  values recognized by the pg_array, pg_hstore, and pg_json extensions
+  work well if the defaults_setter :cache option is also used.
+
+= Other Improvements
+
+* The modification_detection plugin no longer breaks column change
+  detection for new objects.
+
+* Database#copy_table in the postgres adapter now handles errors that
+  occur when processing rows.  Previously, an exception could be
+  raised on the next query in that case.
+
+* The results of the changed_columns method are now cached in many
+  places internally where they are called in a loop.  This results
+  in better performance, especially if the modification_detection or
+  serialization_modification_detection plugins are used.
+
+= Backwards Compatibility
+
+* The pg_interval extension now sets a :ruby_default schema entry for
+  recognized interval defaults to the same value Sequel would return
+  if the default value was returned.  Previously, Sequel would use a
+  string in the :ruby_schema schema value.
+
+* String values in hashes returned by Database#schema are now frozen
+  to prevent possible thread-safety issues and issues with
+  unintentional modification of a shared string.  The hashes
+  themselves are not frozen and can still be modified.

data/doc/release_notes/5.6.0.txt

@@ -0,0 +1,31 @@
+= Improvements
+
+* Running migrations using one of the included migrators on separate
+  Database objects in separate threads simultaneously is now
+  supported.  Previously, the migrators were not thread-safe.
+
+* On Ruby 2.5+, :db_type entries in the schema hashes are now deduped
+  for a slight memory savings when using many columns with the same
+  database type.
+
+* The schema_caching extension now freezes string values in the
+  resulting hashes, just as the default schema parsing code started
+  doing in 5.5.0.
+
+* The schema_caching extension now supports the :callable_default
+  schema values used by the pg_json, pg_array, and pg_hstore
+  extensions, by removing the entry before caching and resetting it
+  after restoring the cache.
+
+* Identifier mangling rules are now respected when renaming columns on
+  Microsoft SQL Server.
+
+= Backwards Compatibility
+
+* The migrator internals were modified in order to support
+  thread-safety.  The private Migrator#remove_migration_classes
+  method has been removed, and #load_migration_file now returns the
+  migration object/class instead of populating Migration.descendants.
+  Migration.descendants is now only used for temporary storage, and
+  will no longer contain all migration objects/classes used by the
+  migrator.

data/doc/release_notes/5.7.0.txt

@@ -0,0 +1,108 @@
+= New Features
+
+* An integer64 extension has been added, which treats the Integer
+  class as a generic 64-bit integer type.  Sequel's default behavior
+  for Integer is to use the integer type, which on most databases
+  is a 32-bit type.
+
+  This affects all internal use of the Integer class as a generic
+  database type, so that methods like primary_key and foreign_key
+  also default to using a 64-bit integer type when using this
+  extension.
+
+* When using PostgreSQL 10+, you can use the :identity option when
+  creating columns to create identity columns:
+
+    DB.create_table(:table){Integer :id, identity: true}
+    # CREATE TABLE "table" ("id" integer GENERATED BY DEFAULT AS IDENTITY)
+
+  If you want to disallow using a user provided value when inserting,
+  or updating you can use a value of :always:
+
+    DB.create_table(:table){Integer :id, identity: :always}
+    # CREATE TABLE "table" ("id" integer GENERATED ALWAYS AS IDENTITY)
+
+* Database#convert_serial_to_identity has been added on PostgreSQL 10.2+.
+  This method can convert existing serial columns to identity columns
+  in most cases, but it currently requires superuser permissions as it
+  modifies the system tables directly.
+
+* Dataset#overriding_system_value and #overriding_user_value are
+  now supported on PostgreSQL to work with identity columns.  You can
+  use #overriding_system_value to force the use of a user provided
+  value for identity columns that are GENERATED ALWAYS, and you can
+  use #overriding_user_value to ignore any user value for identity
+  columns and always use the next entry in the sequence.
+
+= Other Improvements
+
+* On PostgreSQL 10.2+, identity columns are now used instead of serial
+  columns as the default for auto incrementing primary keys:
+
+    DB.create_table(:table){primary_key :id}
+
+    # Sequel 5.7.0+ and PostgreSQL 10.2+
+    # CREATE TABLE "table" ("id" integer
+    #   GENERATED BY DEFAULT AS IDENTITY PRIMARY KEY)
+
+    # Older Sequel version or older PostgreSQL version
+    # CREATE TABLE "table" ("id" serial PRIMARY KEY)
+
+  Identity columns fix many issues that serial columns have, in
+  addition to being the SQL standard way to support auto incrementing
+  columns.
+
+* PostgreSQL identity columns are now correctly recognized and the
+  :auto_increment schema entry is now populated for them.
+
+* Dataset#with_sql_{all,each,first,single_value} now use a cached
+  dataset to avoid clobbering the current dataset's columns.
+  Previously, the clobbering of the current dataset's columns was
+  documented and the method warned against using SQL with
+  different columns.  These methods are now safe to use in such
+  cases, but will not have the same performance advantages if the
+  current dataset is not cached.
+
+* On ruby 2.1+, Sequel now uses Process::CLOCK_MONOTONIC when
+  performing elapsed time calculations so that it is not affected by
+  modifications to the system's time.
+
+* In the postgres adapter, prepared statement errors related to
+  changing types are now treated as disconnect errors. While they
+  are not technically disconnect errors, treating them as such
+  will in general reduce the total number of exceptions generated
+  from 1 per affected statement per connection to 1 per
+  connection.
+
+* In the pg_array_associations plugin, the array_type for
+  pg_array_to_many and many_to_pg_array association reflections is
+  now always the scalar type for the array (e.g. integer). Previously,
+  the array type (e.g. integer[]) was used in some cases.  This didn't
+  previously result in issues as PostgreSQL considers integer[][] the
+  same type as integer[].
+
+* In the pg_array_associations plugin, the many_to_pg_array
+  association remove_all_* method now uses the appropriate cast to
+  work for non-integer array types such as bigint[].
+
+* Database#server_version on PostgreSQL 10.1+ now works correctly
+  when the connection does not support the server_version method.
+  Now the server_version_num database setting is always used to
+  ensure consistent behavior across adapters.
+
+* In the jdbc/oracle adapter, temporary clobs are now manually
+  freed to prevent a memory leak, in line with the Oracle JDBC
+  driver recommendations.
+
+* The Sequel <4 release notes and changelog are no longer shipped
+  with the gem, decreasing the size of the gem by 20%.
+
+= Backwards Compatibility
+
+* The switch to using identity columns instead of serial columns
+  by default on PostgreSQL 10.2+ may break backwards compatibilty
+  in some situations, such as code that relies on what are generally
+  considered bugs in serial columns, such as CREATE TABLE LIKE
+  using the same sequence for the column in both the existing table
+  and the new table, or that dropping the default value for the
+  column does not drop the related sequence.

data/doc/release_notes/5.8.0.txt

@@ -0,0 +1,170 @@
+= New Features
+
+* A pg_auto_constraint_validations plugin has been added, which
+  automatically converts many constraint violations raised as
+  exceptions to ValidationFailed exceptions when saving a model
+  instance.
+
+  The following constraint violation types are recognized and
+  supported:
+
+  * NOT NULL
+  * CHECK
+  * UNIQUE (except expression/functional indexes)
+  * FOREIGN KEY (both referencing and referenced by)
+
+  In the cases where the plugin cannot determine an appropriate
+  validation failure for the constraint violation, it just
+  reraises the original exception.
+
+  This plugin is not intended as a replacement for other validations,
+  it is intended as a last resort.  The purpose of validations is to
+  provide nice error messages for the user, and the error messages
+  generated by this plugin are fairly generic.  The error messages can
+  be customized using the :messages plugin option, but there is only a
+  single message used per constraint type.
+
+* Database#check_constraints has been added on PostgreSQL.  This
+  returns metadata related to each check constraint on a table:
+
+    DB.create_table(:foo) do
+      Integer :i
+      Integer :j
+      constraint(:ic, Sequel[:i] > 2)
+      constraint(:jc, Sequel[:j] > 2)
+      constraint(:ijc, Sequel[:i] - Sequel[:j] > 2)
+    end
+    DB.check_constraints(:foo)
+    # => {
+    #  :ic=>{:definition=>"CHECK ((i > 2))", :columns=>[:i]},
+    #  :jc=>{:definition=>"CHECK ((j > 2))", :columns=>[:j]},
+    #  :ijc=>{:definition=>"CHECK (((i - j) > 2))", :columns=>[:i, :j]}
+    # }
+
+* Database#foreign_key_list now supports a :reverse option on
+  PostgreSQL, which returns foreign keys referencing the given table,
+  instead of of foreign keys in the given table referencing other
+  tables:
+
+    DB.create_table!(:a) do
+      primary_key :id
+      Integer :i
+      Integer :j
+      foreign_key :a_id, :a, :foreign_key_constraint_name=>:a_a
+      unique [:i, :j]
+    end
+    DB.create_table!(:b) do
+      foreign_key :a_id, :a, :foreign_key_constraint_name=>:a_a
+      Integer :c
+      Integer :d
+      foreign_key [:c, :d], :a, :key=>[:j, :i], :name=>:a_c_d
+    end
+    DB.foreign_key_list(:a, :reverse=>true)
+    # => [
+    #  {:name=>:a_a, :columns=>[:a_id], :key=>[:id], :on_update=>:no_action,
+    #   :on_delete=>:no_action, :deferrable=>false, :table=>:a, :schema=>:public},
+    #  {:name=>:a_a, :columns=>[:a_id], :key=>[:id], :on_update=>:no_action,
+    #   :on_delete=>:no_action, :deferrable=>false, :table=>:b, :schema=>:public},
+    #  {:name=>:a_c_d, :columns=>[:c, :d], :key=>[:j, :i], :on_update=>:no_action,
+    #   :on_delete=>:no_action, :deferrable=>false, :table=>:b, :schema=>:public}
+    # ]
+
+* Dataset#nowait has been added, which will make the query fail
+  with a Sequel::DatabaseLockTimeout exception if it encounters
+  a locked row, overriding the default database behavior that
+  would wait until the lock was released. This method is supported
+  on PostgreSQL, Microsoft SQL Server, Oracle, and MySQL 8+.
+
+* Database#indexes now supports an :include_partial option on
+  PostgreSQL, which will include partial indexes in the output (Sequel
+  by default excludes partial indexes).
+
+* Common table expressions and window functions are now supported when
+  using MySQL 8+.
+
+* Dataset#skip_locked is now supported on MySQL 8+.
+
+* The connection_expiration extension now supports a
+  Database#connection_expiration_random_delay attribute, which is used
+  to randomize the expiration times, avoiding the thundering herd
+  problem.
+
+* The pg_enum extension now supports a rename_enum method for renaming
+  existing enum types.
+
+* Database#error_info on PostgreSQL now returns much more metadata
+  regarding the error.
+
+= Other Improvements
+
+* The dataset returned by the following dataset methods is cached,
+  which can improve performance significantly in certain cases:
+
+  * #distinct (without arguments or block)
+  * #from_self (without options)
+  * #lateral
+  * #qualify (without argument)
+  * #returning (without arguments)
+  * #select_all (without arguments)
+
+* If the primary_key serial: true, type: :serial, or type: :bigserial
+  options are given on PostgreSQL 10.2+, use a serial primary key
+  instead of an identity primary key.  This change was included in
+  Sequel 5.7.1.
+
+* The :search_path Database option is now supported as a shard
+  option on PostgreSQL, so different shards can use different
+  search paths.
+
+* The correct column order in Database#foreign_key_list on MySQL is
+  now forced, fixing issues on MySQL 8+.
+
+* When using case sensitive regexp matches on MySQL 8+, Sequel now
+  uses the REGEXP_LIKE function instead of the REGEXP BINARY
+  operator, to work around what appears to be a bug in MySQL 8+
+  related to the change in MySQL's regexp engine.
+
+* On MySQL 5.7+, the :extended option to Dataset#explain is now
+  ignored, since the :extended option's behavior in previous
+  MySQL versions is now the default behavior.
+
+* The MySQL HY000 generic SQL state error code is now ignored
+  in the mysql2 adapter, so it falls back to using the more
+  accurate backup error mapping in that case.
+
+* The pg_enum extension's schema modification methods now work
+  correctly if the Database instance is frozen.
+
+* The tactical_eager_loading plugin now respects the :allow_eager
+  association option, and will not attempt to eagerly load
+  associations when :allow_eager is false.
+
+* Using multiple add_constraint calls and a set_column_null call in
+  the same alter_table block on SQLite now works correctly.  Note
+  that if you are planning on ever modifying existing tables beyond
+  adding columns, you should probably choose a database that natively
+  supports such modification (SQLite does not).
+
+* Hashes returned by Database#foreign_key_list on PostgreSQL now
+  include a :schema entry, unless the support has been enabled
+  to make the :table entry be a qualified identifier.
+
+* Dataset#support_cte?(:insert) no longer returns true on
+  SQLAnywhere.  SQLAnywhere only supports common table
+  expressions for INSERT ... SELECT, not for all INSERT
+  statements. INSERT ... WITH ... SELECT is already
+  supported in Sequel using:
+
+    DB[:t1].insert(DB[:t2].with(DB[:t3]))
+
+* Model#_valid? is no longer made a public method in the
+  error_splitter plugin.
+
+= Backwards Compatibility
+
+* Calling the filter method on a proxy object returned by the
+  association_proxies plugin now warns on ruby <2.6.  This is
+  because starting in ruby 2.6, the behavior will change and the
+  method will be called on the array of associated objects
+  instead of on the dataset, as Enumerable#filter is being added
+  in ruby 2.6.

data/doc/release_notes/5.9.0.txt

@@ -0,0 +1,99 @@
+= New Features
+
+* An escaped_like extension has been added, for the creation of
+  LIKE/ILIKE expressions with placeholders in patterns without
+  access to a dataset.  This adds escaped_like and escaped_ilike
+  methods to the same Sequel expression objects that support like
+  and ilike.  These methods take two arguments, the first being
+  the pattern, with ? placeholders, and the second being the
+  placeholder value (which can be an array for multiple
+  placeholders):
+
+    Sequel.extension :escaped_like
+    DB[:table].where{string_column.escaped_like('?%', user_input)}
+    # user_input is 'foo':
+    #  SELECT * FROM table WHERE string_column LIKE 'foo%'
+    # user_input is '%foo':
+    #  SELECT * FROM table WHERE string_column LIKE '\%foo%'
+
+* Generated columns on MySQL 5.7+ and MariaDB 5.2+ are now supported
+  using the :generated_always_as option when creating the column.
+  The :generated_type option can also be used to specify the type of
+  generated column (virtual or stored).  Examples:
+
+    DB.add_column :t, :c, Integer, generated_always_as: Sequel[:a]+'b'
+    # ALTER TABLE `t` ADD COLUMN `c` varchar(255)
+    # GENERATED ALWAYS AS (CONCAT(`a`, 'b'))
+
+    DB.add_column :t, :c, Integer, generated_always_as: Sequel[:a]+'b',
+      generated_type: :virtual
+    # ALTER TABLE `t` ADD COLUMN `c` varchar(255)
+    # GENERATED ALWAYS AS (CONCAT(`a`, 'b')) VIRTUAL
+
+    DB.add_column :t, :c, Integer, generated_always_as: Sequel[:a]+'b',
+      generated_type: :stored
+    # ALTER TABLE `t` ADD COLUMN `c` varchar(255)
+    # GENERATED ALWAYS AS (CONCAT(`a`, 'b')) STORED
+
+* Sequel::Model.has_dataset? has been added for checking whether the
+  model class has an associated dataset.  This will generally be true
+  for most model classes, but will be false for abstract model
+  classes (such as Sequel::Model itself).
+
+* Sequel::VERSION_NUMBER has been added for easier future version
+  comparisons.  The version number for 5.9.0 is 50090.
+
+= Other Improvements
+
+* When disconnecting connections in the threaded connection pools,
+  the disconnection is performed without holding the connection
+  pool mutex, since disconnection may block.
+
+* The sharded threaded connection pool no longer deadlocks when
+  disconnecting connections if the connection_validator or
+  connection_expiration extension is used.
+
+* If a thread dies and does not check a connection back into the
+  connection pool, Sequel now disconnects the connection when it
+  detects the dead thread, instead of assuming the connection is
+  safe to be reused.
+
+* When using eager_graph with cascaded associations, a unique
+  object is now used instead of a shared object in cases where
+  using a shared object may cause further cascaded associated
+  objects to be duplicated.
+
+* On PostgreSQL, the ESCAPE modifier to the LIKE/ILIKE operators is
+  no longer used, since the default ESCAPE value is the one Sequel
+  uses.  This change was made in order to allow the LIKE/ILIKE
+  operators to work with the ANY function, as PostgreSQL does not
+  support the use of the ESCAPE modifier in such cases.
+
+* A hash argument passed to Model.nested_attributes in the
+  nested_attributes plugin is now no longer modified.
+
+* Internal data structures for eager and eager_graph datasets are now
+  frozen to avoid unintentional modification.
+
+* Nondeterministic behavior in Database#foreign_key_list with the
+  :reverse option on PostgreSQL is now avoided by using an
+  unambiguous order.
+
+* Performance has been improved slightly by avoiding unnecessary
+  hash allocations.
+
+* Performance has been improved slightly by using while instead
+  of Kernel#loop.
+
+* BigDecimal() is now used instead of BigDecimal.new(), as the
+  latter has been deprecated.
+
+* The jdbc adapter now avoids referencing ::NativeException on JRuby
+  9.2+, since JRuby has deprecated it.  It is still used on older
+  versions of JRuby, since some JRuby 1.7 code may still require it.
+
+* Sequel now works around multiple Date/Time conversion bugs in
+  JRuby 9.2.0.0 for BC dates in the pg_extended_date_support
+  extension.  These bugs have already been fixed in JRuby, and
+  the workarounds will be removed after the release of JRuby
+  9.2.1.0.