sequel 4.36.0 → 5.61.0

data/doc/release_notes/4.23.0.txt

@@ -1,65 +0,0 @@
-= New Features
-
-* An update_refresh plugin has been added, for refreshing a model
-  instance when updating.  The default behavior is to only refresh
-  when inserting.  However, if you have triggers on the model's table,
-  it's a good idea to refresh when updating to pick up the possibly
-  changed values.  On databases that support UPDATE RETURNING, such as
-  PostgreSQL, the update and refresh are done in a single query.
-
-* A delay_add_association plugin has been added, for delaying add_*
-  method calls for associations until after the receiver has been
-  saved, if the receiver is a new object. Example:
-
-    artist = Artist.new(:name=>'Foo')
-    artist.add_album(Album.new(:name=>'Bar'))
-    # No database queries yet
-
-    artist.save # Saves artist, then album
-
-* A validate_associated plugin has been added, for validating
-  associated objects when validating the current object.  This
-  was extracted from the nested_attributes plugin, and is also
-  used by the delay_add_association plugin.  For example,
-  if you have an albums association and you want to validate all
-  associated objects before saving the current object, you can
-  cal validate_associated_object for each object:
-
-    def validate
-      super
-      reflection = association_reflection(:albums)
-      associations[:albums].each do |obj|
-        validate_associated_object(reflection, obj)
-      end
-    end
-
-= Other Improvements
-
-* Database#transaction now returns the block return value if
-  :rollback=>:always is used.  Previously, it would return nil in
-  that case.
-
-* Postgres::JSONBOp#[] and #get_text now return JSONBOp instances
-  instead of JSONOp instances.
-
-* Model#move_to, #move_up, and #move_down in the list plugin now
-  automatically handle out-of-range targets by defaulting to the first
-  or last position in the list.  Previously, using an out of range
-  target would raise an exception.
-
-* Database#add_named_conversion_proc on PostgreSQL now works for enum
-  types.
-
-* dataset.call_sproc(:insert, ...) now works correctly on JDBC.
-
-* postgresql:// connection strings are now supported, since that is
-  the protocol name supported by libpq.
-
-* Sequel has switched from rspec to minitest/spec for testing, and
-  now uses random test order when testing. During the conversion
-  process, many test order dependency bugs were fixed.
-
-= Backwards Compatibility
-
-* The deprecated fdbsql, jdbc/fdbsql, and openbase adapters have been
-  removed.

data/doc/release_notes/4.24.0.txt

@@ -1,99 +0,0 @@
-= New Features
-
-* A pg_inet_ops extension has been added, for DSL support for
-  calling PostgreSQL inet functions and operators.  Example:
-
-    r = Sequel.pg_inet_op(:inet)
-
-    ~r                                 # ~inet
-    r & :other                         # inet & other
-    r | :other                         # inet | other
-    r << :other                        # inet << other
-    r >> :other                        # inet >> other
- 
-    r.contained_by(:other)             # inet << other
-    r.contained_by_or_equals(:other)   # inet <<= other
-    r.contains(:other)                 # inet >> other
-    r.contains_or_equals(:other)       # inet >>= other
-    r.contains_or_contained_by(:other) # inet && other
- 
-    r.abbrev           # abbrev(inet)
-    r.broadcast        # broadcast(inet)
-    r.family           # family(inet)
-    r.host             # host(inet)
-    r.hostmask         # hostmask(inet)
-    r.masklen          # masklen(inet)
-    r.netmask          # netmask(inet)
-    r.network          # network(inet)
-    r.set_masklen(16)  # set_masklen(inet, 16)
-    r.text             # text(inet)
-
-* The association_pks plugin now supports a :delay_pks association
-  option.  When set to true, this makes the methods created by the
-  plugin usable on new objects, by delaying the saving of the
-  associated pks until after the new object has been saved.  When
-  set to :always, this also changes the behavior of the methods
-  for existing objects, so that nothing is persisted until the
-  object has been saved.  Example:
-
-    Album.plugin :association_pks
-    Album.many_to_many :tags, :delay_pks=>true
-
-    album = Album.new(:tag_pks=>[1,2,3]) # No database query
-    album.save # Queries to insert album, and then update albums_tags
-
-* The class_table_inheritance plugin now supports subclasses that
-  don't require additional columns, and therefore do not need to
-  join to additional tables.  It now loads the
-  single_table_inheritance plugin and supports options that were
-  previously only supported by single_table_inheritance, such as the
-  :key_map and :key_chooser options.
-
-* The validation_helpers plugin now supports a :from=>:values option
-  in the validation methods, which will take the value directly from
-  the values hash instead of calling the related method.  This
-  allows validation_helpers to differentiate between validations on
-  underlying database column and validations on the model.
-
-  The auto_validations plugin has been modified to use this feature,
-  since all validations it generates are for validations on the
-  underlying database columns.
-
-* The auto_validations plugin now supports options to pass to each
-  of the underlying validation methods:
-
-    Sequel::Model.plugin :auto_validations,
-      :unique_opts=>{:only_if_modified=>true}
-
-  In addition to :unique_opts, there is support for :not_null_opts
-  (for NOT NULL columns without a default), :explicit_not_null_opts
-  (for NOT NULL columns with a default), :max_length_opts, and
-  :schema_types_opts. 
-
-* The update_refresh plugin now accepts a :columns option, which
-  specifies the columns to refresh.  This option is currently only
-  respected if the related dataset supports RETURNING.
-
-* The :timeout option to Database#listen in the postgres adapter can
-  now be a callable object, previously it had to be Numeric.  This
-  allows you to dynamically change the timeout based on current
-  application state.
-
-= Other Improvements
-
-* The uniqueness validations added by the auto_validations plugin now
-  use a symbol key in the related Errors instance if the underlying
-  index was on a single column.  Previously, the uniqueness
-  validations for a single column would use an array key in the
-  related Errors instance.
-
-* The jdbc subadapters now correctly handle 64-bit autoincrementing
-  primary keys.
-
-* The jdbc subadapters now work correctly if they issue queries while
-  the subadapter is being loaded.  This can happen in the
-  jdbc/postgresql adapter if the pg_hstore extension is used.
-
-= Backwards Compatibility
-
-* The deprecated db2 and dbi adapters have been removed.

data/doc/release_notes/4.25.0.txt

@@ -1,181 +0,0 @@
-= New Features
-
-* The =~ and !~ methods are now defined on ComplexExpressions in
-  addition to GenericExpressions, allowing the following code to
-  work:
-
-    DB[:table].where{(column1 + column2) =~ column3}
-
-* Dataset#group_append has been added for appending to an existing
-  GROUP BY clause:
-
-    ds = DB[:table].group(:column1)
-    # SELECT * FROM table GROUP BY column1
-
-    ds = ds.group_append(:column2)
-    # SELECT * FROM table GROUP BY column1, column2
-
-* An inverted_subsets plugin has been added, for automatic creation of
-  methods for the inversion of the subset criteria.  For example:
-
-    Album.plugin :inverted_subsets
-    Album.subset :published, :published=>true
-
-    Album.published
-    # SELECT * FROM albums WHERE published IS TRUE
-
-    Album.not_published
-    # SELECT * FROM albums WHERE published IS NOT TRUE
-
-  By default, the subset method name is prefixed with "not_".  You can
-  pass a block to override the default behavior:
-
-    Album.plugin(:inverted_subsets){|name| "exclude_#{name}"}
-    Album.subset :published, :published=>true
-
-    Album.exclude_published
-    # SELECT * FROM albums WHERE published IS NOT TRUE
-
-* A singular_table_names plugin has been added, which changes Sequel
-  to not pluralize table names by default.
-
-    Sequel::Model.plugin :singular_table_names
-    class FooBar < Sequel::Model; end
-    FooBar.table_name # => foo_bar
-
-* Dataset#insert_conflict and #insert_ignore have been added on
-  PostgreSQL.  When using PostgreSQL 9.5+, they allow you to ignore
-  unique or exclusion constraint violations on inserting, or to do
-  an update instead:
-
-    DB[:table].insert_conflict.insert(:a=>1, :b=>2)
-    # INSERT INTO TABLE (a, b) VALUES (1, 2)
-    # ON CONFLICT DO NOTHING
-
-  You can pass a specific constraint name using :constraint, to only
-  ignore a specific constraint violation:
-
-    DB[:table].insert_conflict(:constraint=>:table_a_uidx).
-      insert(:a=>1, :b=>2)
-    # INSERT INTO TABLE (a, b) VALUES (1, 2)
-    # ON CONFLICT ON CONSTRAINT table_a_uidx DO NOTHING
-   
-  If the unique or exclusion constraint covers the whole table (e.g.
-  it isn't a partial unique index), then you can just specify the
-  column using the :target option:
-
-    DB[:table].insert_conflict(:target=>:a).insert(:a=>1, :b=>2)
-    # INSERT INTO TABLE (a, b) VALUES (1, 2)
-    # ON CONFLICT (a) DO NOTHING
-   
-  If you want to update the existing row instead of ignoring the
-  constraint violation, you can pass an :update option with a hash of
-  values to update.  You must pass either the :target or :constraint
-  options when passing the :update option:
-
-    DB[:table].insert_conflict(:target=>:a,
-                               :update=>{:b=>:excluded__b}).
-      insert(:a=>1, :b=>2)
-    # INSERT INTO TABLE (a, b) VALUES (1, 2)
-    # ON CONFLICT (a) DO UPDATE SET b = excluded.b
-   
-  Additionally, if you only want to do the update in certain cases,
-  you can specify an :update_where option, which will be used as a
-  filter.  If the row doesn't match the conditions, the constraint
-  violation will be ignored, but the row will not be updated:
-
-    DB[:table].insert_conflict(:constraint=>:table_a_uidx,
-                               :update=>{:b=>:excluded__b},
-                               :update_where=>{:table__status_id=>1}).
-      insert(:a=>1, :b=>2)
-    # INSERT INTO TABLE (a, b) VALUES (1, 2)
-    # ON CONFLICT ON CONSTRAINT table_a_uidx
-    # DO UPDATE SET b = excluded.b WHERE (table.status_id = 1)
-
-* Dataset#group_rollup and #group_cube are now supported when using
-  PostgreSQL 9.5+.
-
-* Sequel now supports Dataset#returning when using prepared statements
-  and bound variables:
-
-    DB[:table].returning.prepare(:insert, :i, :col=>:$col).
-      call(:col=>42)
-    # => [{:col=>42}]
-
-= Other Improvements
-
-* The serialization plugin now integrates with the dirty plugin, so
-  that column changes are detected correctly.  However, column values
-  that are changed and then changed back to the original value are
-  still detected as changed.
-
-* Dataset#for_update and similar locking methods now cause Sequel not
-  to use the :read_only shard if sharding is used.  
-
-* The association_pks plugin now clears cached delayed associated pks
-  when the object is refreshed.
-
-* The :collate column option when adding columns now literalizes
-  non-String values on PostgreSQL. Previously, the :collate option
-  value was used verbatim. This is because PostgreSQL's collations
-  generally require quoting as they are uppercase or mixed-case.
-
-* Sequel's metadata parsing methods now support Microsoft SQL Server
-  2012+ when used in case sensitive mode.
-
-* Sequel now recognizes an addition check constraint violation
-  exception on SQLite.
-
-* Sequel now recognizes constraint violations when using the
-  swift/sqlite adapter.
-
-* Sequel now automatically REORGs tables when altering them in the
-  jdbc/db2 adapter.
-
-= Backwards Compatibility
-
-* Sequel now defaults to ignoring NULL values when using IN/NOT IN
-  with an empty array.  Previously, code such as:
-
-    DB[:table].where(:column=>[])
-
-  would be literalized as:
-
-    SELECT * FROM table WHERE (column != column)
-
-  This yields a NULL value when column is NULL, similarly to how most
-  other SQL operators work.  Unfortunately, most databases do not
-  optimize this, and such a query can require a sequential scan of the
-  table.
-
-  Sequel previously shipped with a empty_array_ignore_nulls extension
-  that literalized the query to:
-
-    SELECT * FROM table WHERE (1 = 0)
-
-  which databases will generally optimize to a constant false value,
-  resulting in much faster queries.  This behavior is now the default.
-
-  Users that desire the previous behavior can use the new
-  empty_array_consider_nulls extension.
-
-* The deprecated firebird and informix adapters have been removed.
-
-* Calling prepare on a prepared statement now raises an exception.
-  It was supported accidently before, as prepared statements are
-  dataset instances.
-
-* Model::DatasetModule#subset now calls Model.subset instead of
-  the other way around.  This makes it possible to modify the
-  behavior of subset in a plugin.
-
-* The :collate column option change on PostgreSQL can break code
-  that used already quoted values in symbols.  For example:
-
-    String :column_name, collate=>:'"C"'
-
-  would need to change to:
-
-    String :column_name, collate=>:C
-    # or
-    String :column_name, collate=>'"C"'

data/doc/release_notes/4.26.0.txt

@@ -1,44 +0,0 @@
-= New Features
-
-* Add Dataset#grouping_sets to support GROUP BY GROUPING SETS on
-  PostgreSQL 9.5+, MSSQL 2008+, Oracle, DB2, and SQLAnywhere:
-
-    DB[:test].group([:type_id, :b], :type_id, []).grouping_sets
-    # SELECT * FROM test
-    # GROUP BY GROUPING SETS((type_id, b), (type_id), ())
-
-* Sequel::NoMatchingRow exceptions raised by Sequel now give access
-  to the dataset that raised the exception via the dataset method.
-  This makes it easier to write generic error handling code.
-
-* Support :if_exists option to drop_column on PostgreSQL:
-
-    DB.drop_column :t, :col, :if_exists=>true 
-    ALTER TABLE t DROP COLUMN IF EXISTS col
-
-= Other Improvements
-
-* Make the class_table_inheritance plugin work correctly without an
-  sti_key.  This was broken in a recent refactoring to make class
-  table inheritance support multiple classes for a single table.
-
-* Make Class.new(ModelClass){set_dataset :table} work correctly on
-  ruby 1.8.  This was broken in a refactoring to allow the
-  singular_table_names plugin to work.
-
-* Make offset emulation via ROW_NUMBER better handle ambiguous column
-  names for datasets without an ORDER BY clause, but with an explicit
-  SELECT clause.
-
-* Make pg_range extension use PostgreSQL range function constructors
-  instead of casting string literals to the appropriate range type,
-  if the range type is known.  This allows arbitrary expressions to
-  be used inside ranges, such as CURRENT_TIMESTAMP in timestamp
-  ranges.
-
-* Make Dataset#== not consider frozen status.
-
-* Allow Dataset#prepare on already prepared statements in situations
-  where determining the SQL for a prepared statement requires it.
-
-* Detect additional disconnect errors when using the tinytds adapter.

data/doc/release_notes/4.27.0.txt

@@ -1,78 +0,0 @@
-= New Features
-
-* A before_after_save plugin has been added, which for newly
-  created objects refreshes the object before calling after_create,
-  and resets the modified flag before calling after_update.
-  Previously, these actions were not taken until after after_save
-  was called.  This will be the default behavior in Sequel 5.
-
-* In create_table blocks, primary_key now supports a :keep_order
-  option, which will not change the order in which the primary key
-  is added.  Without this option, Sequel's historical behavior of
-  making the primary key column the first column is used.
-
-    DB.create_table(:foo) do
-      Integer :a
-      primary_key :b, :keep_order=>true
-    end
-    # CREATE TABLE foo 
-    # (a integer, b integer PRIMARY KEY AUTOINCREMENT)
-
-  The schema dumper now uses this option if necessary, allowing it
-  to correctly dump tables where the primary key column is not the
-  first column.
-
-* Dataset#single_record! and #single_value! have been added.  These
-  are faster versions of #single_record and #single_value that
-  don't require cloning the dataset.  If you are sure the dataset
-  will only return a single row or a single value, you can use
-  these methods for better performance.
-
-* The new jsonb and json functions added in PostgreSQL 9.5 are now
-  supported by the pg_json_ops extension.
-
-    Sequel.pg_jsonb_op(:metadata).set(%w'a b', [1,2,3])
-    # jsonb_set("metadata", ARRAY['a','b'], '[1,2,3]'::jsonb, true)
-
-= Other Improvements
-
-* Sequel.synchronize is no longer a stub on MRI.  Testing has shown
-  that relying on the global interpreter lock to protect
-  multi-threaded access to hashes is not safe in all environments,
-  so Sequel now uses a mutex on MRI just as it does on other ruby
-  interpreters.
-
-* Database#schema now sets the :auto_increment option correctly for
-  auto incrementing primary keys if they are not the first column
-  in the table.
-
-* Dataset#single_value and #with_sql_single_value are now slightly
-  faster by avoiding an array allocation.
-
-* Model datasets can now use #with_sql_single_value and return a
-  single value, instead of an array in [:column_name, value] format.
-
-* Model#persisted? in the active_model plugin will now return false
-  if the transaction that inserts the row for the object is rolled
-  back.
-  
-* bin/sequel now warns if additional arguments are passed that it
-  ignores.  In Sequel 5, bin/sequel will raise an error in these
-  cases.
-
-* Database#foreign_key_list on PostgreSQL now returns referenced
-  composite keys in the correct order.
-
-* The postgres adapter now works with postgres-pr 0.7.0. Note that
-  postgres adapter users that want a pure-ruby driver are encouraged
-  to use jeremyevans-postgres-pr as that has many additional bugfixes
-  and is the version tested with Sequel on a regular basis.
-
-* The jdbc/postgresql adapter now recognizes an additional disconnect
-  error.
-
-= Backwards Compatibility
-
-* Users who were relying on #with_sql_single_value returning an array
-  instead of a single value for model datasets need to update their
-  code.

data/doc/release_notes/4.28.0.txt

@@ -1,57 +0,0 @@
-= New Features
-
-* A subset_conditions plugin has been added, which adds a method
-  for each subset that returns the filter conditions for the
-  subset.  This makes it easier to reuse the subset conditions:
-
-    class Foo < Sequel::Model
-      plugin :subset_conditions
-      subset :active, :active=>true
-    end
-
-    Foo.exclude(Foo.active_conditions)
-    Foo.where(:a=>1).or(Foo.active_conditions)
-
-* A boolean_subsets plugin has been added, which adds a subset for each
-  boolean column:
-
-    # Assume boolean column :active
-    Foo.plugin :boolean_subsets
-
-    Foo.active
-    # SELECT * FROM foos WHERE (active IS TRUE)
-
-  You can provide a block to the plugin to change the arguments passed
-  to subset:
-    
-    Foo.plugin :boolean_subsets do |column|
-      [:"where_#{column}", column]
-    end
-
-    Foo.where_active
-    # SELECT * FROM foos WHERE active
-
-  As with similar plugins, you can add the boolean_subsets plugin to
-  Sequel::Model itself, and all subclasses created afterward will have
-  the boolean subset methods automatically created.
-
-= Other Improvements
-
-* If Model#refresh can't find the related row, Sequel now raises a
-  Sequel::NoExistingObject exception instead of a generic
-  Sequel::Error exception.
-
-* In the csv_serializer plugin, when calling #to_csv on a model class
-  or dataset, instead of using #[] to access data, #send is used to
-  call methods.  This is more similar to other plugins as well as
-  Model#to_csv.
-
-* The list plugin now works better with the auto_validations plugin,
-  or any other time there is a validation on the position column.
-
-= Backwards Compatibility
-
-* The change to the csv_serializer plugin can change results if you
-  are overriding any of the column accessor methods.  It can also
-  break existing code if one of the columns being used isn't defined
-  as a method or the method requires more than one argument.

data/doc/release_notes/4.29.0.txt

@@ -1,41 +0,0 @@
-= New Features
-
-* A uuid plugin has now been added.  This plugin will automatically
-  create a uuid for newly created model objects.
-
-    Model.plugin :uuid
-    Model.create.uuid => # some UUID
-
-* Model#json_serializer_opts has been added to the json_serializer
-  plugin, allowing you to override the JSON serialization options
-  on a per instance basis without passing the options directly
-  to Model#to_json.  This is useful if you are including the model
-  instance inside another datastructure that will be serialized
-  to JSON.
-
-    obj.json_serializer_opts(:root => true)
-    [obj].to_json
-    # => '[{"obj":{"id":1,"name":"Foo"}}]'
-
-= Other Improvements
-
-* The Database#transaction :retry_on option now works when using
-  savepoints.
-
-* Calling Database#table_exists? inside a transaction will now use
-  a savepoint if the database supports it, so that if the table
-  doesn't exist, it will not affect the state of the transaction.
-
-* Blobs can now be used as bound variables in the oracle adapter.
-
-* The sqlanywhere adapter now works with database sharding.
-
-* The Dataset#full_text_search :rank option has been fixed to order
-  by rank descending instead of ascending.
-
-* External adapters that do not support INSERT with DEFAULT VALUES
-  can now override Dataset#insert_empty_columns_values to set
-  the columns and values to use for an empty INSERT.
-
-* External adapters can now implement Dataset#date_add_sql_append
-  to integrate with the date_arithmetic extension.

data/doc/release_notes/4.3.0.txt

@@ -1,40 +0,0 @@
-= New Features
-
-* The tree and rcte_tree plugins now support composite keys.
-
-* An error_sql Database extension has been added.  This extension
-  adds the DatabaseError#sql method, which should return the
-  database query that caused the error.  This is useful for
-  drivers that don't include the SQL used as part of the error
-  message.
-
-= Other Improvements
-
-* Empty blobs are now literalized correctly on MySQL.
-
-* Empty arrays are now literalized correctly on PostgreSQL <8.4.
-
-* In the pagination extension, Dataset#page_count is now 1 even if
-  the dataset is empty.  This fixes issues with last_page? and
-  page_range returning bad values for empty datasets.
-
-* In the pagination extension, calling Dataset#each_page without a
-  block now returns an Enumerator.
-
-* Dataset#qualify and Sequel.delay now work together, qualifying
-  the object returned by the delayed evaluation.
-
-* Migrator.migrator_class is now a public method.
-
-* The PostgreSQL citext type is now recognized as a string.
-
-* Another disconnect error is now recognized in the jdbc/as400
-  adapter.
-
-* Guides about using and creating Sequel extensions and model
-  plugins have been added.
-
-= Backwards Compatibility
-
-* If you were expecting Dataset#page_count on a empty paginated
-  dataset to return 0, you need to update your code.

data/doc/release_notes/4.30.0.txt

@@ -1,37 +0,0 @@
-= New Features
-
-* Overriding the :limit and :eager_limit_strategy association options
-  can now be done on a per-call basis when eager loading, by using an
-  eager block callback and setting the :eager_limit or
-  :eager_limit_strategy dataset options. Example:
-
-    Album.eager(:tracks=>proc{|ds| ds.clone(:eager_limit=>5)}).all
-
-* Dataset#insert_conflict and #insert_ignore have been added on
-  SQLite, adding support for the INSERT OR ... SQL syntax:
-
-    DB[:table].insert_ignore.insert(:a=>1, :b=>2)
-    # INSERT OR IGNORE INTO TABLE (a, b) VALUES (1, 2)  
-
-    DB[:table].insert_conflict(:replace).insert(:a=>1, :b=>2)
-    # INSERT OR REPLACE INTO TABLE (a, b) VALUES (1, 2)
-
-* An identifier_columns plugin has been added, which allows
-  Sequel::Model#save to work when column names contain double
-  underscores.
-
-= Other Improvements
-
-* IPv6 addresses can now be used in connection URLs when using
-  ruby 1.9.3+.
-
-* The :db_type entries in column schema hashes now include sizes
-  for string and decimal types on DB2 and when using the jdbc
-  adapter's generic schema parsing.
-
-* Database#row_type in the pg_row extension now handles different
-  formats of specifying schema qualified types.  So a row type
-  registered via :schema__type can be found using
-  Sequel.qualify(:schema, :type).
-
-* Another disconnect error is recognized in the tinytds adapter.