=== HEAD
 
* Enable almost full support of PostgreSQL-JDBC using the JDBC adapter (jeremyevans)
 
* Remove Sequel::Worker (jeremyevans)
 
* Make PostgreSQL adapter not raise an error when inserting records into a table without a primary key (jeremyevans)
 
* Make Database.uri_to_options a private class method (jeremyevans)
 
* Make JDBC load drivers automatically for PostgreSQL, MySQL, SQLite, Oracle, and MSSQL (jeremyevans)
 
* Make Oracle adapter work with a nonstandard Oracle database port (pavel.lukin)
 
* Typecast '' to nil by default for non-string non-blob columns, add typecast_empty_string_to_nil= model class and instance methods (jeremyevans)
 
* Use a simpler select in Dataset#empty?, fixes use with MySQL (jeremyevans)
 
* Add integration test suite, testing sequel against a real database, with nothing mocked (jeremyevans)
 
* Make validates_length_of default tag depend on presence of options passed to it (jeremyevans)
 
* Combine the directory structure for sequel_model and sequel_core, now there is going to be only one gem named sequel (jeremyevans)
 
=== 2.2.0 (2008-07-05)
 
* Add :extend association option, extending the dataset with module(s) (jeremyevans)
 
* Add :after_load association callback option, called after associated objects have been loaded from the database (jeremyevans)
 
* Make validation methods support a :tag option, to work correctly with source reloading (jeremyevans)
 
* Add :before_add, :after_add, :before_remove, :after_remove association callback options (jeremyevans)
 
* Break many_to_one association setter method in two parts, for easier overriding (jeremyevans)
 
* Model.validates_presence_of now considers false as present instead of absent (jeremyevans)
 
* Add Model.raise_on_save_failure, raising errors on save failure instead of return false (now nil), default to true (jeremyevans)
 
* Add :eager_loader association option, to specify code to be run when eager loading (jeremyevans)
 
* Make :many_to_one associations support :dataset, :order, :limit association options, as well as block arguments (jeremyevans)
 
* Add :dataset association option, which overrides the default base dataset to use (jeremyevans)
 
* Add :eager_graph association option, works just like :eager except it uses #eager_graph (jeremyevans)
 
* Add :graph_join_table_join_type association option (jeremyevans)
 
* Add :graph_only_conditions and :graph_join_table_only_conditions association options (jeremyevans)
 
* Add :graph_block and :graph_join_table_block association options (jeremyevans)
 
* Set the model's dataset's columns in addition to the model's columns when loading the schema for a model (jeremyevans)
 
* Make caching work correctly with subclasses (jeremyevans)
 
* Add the Model.to_hash dataset method (jeremyevans)
 
* Filter blocks now yield a SQL::VirtualRow argument, which is useful if another library defines operator methods on Symbol (jeremyevans)
 
* Add Symbol#identifier method, to make x__a be treated as "x__a" instead of "x"."a" (jeremyevans)
 
* Dataset#update no longer takes a block, please use a hash argument with the expression syntax instead (jeremyevans)
 
* ParseTree support has been removed from Sequel (jeremyevans)
 
* Database#drop_column is now supported in the SQLite adapter (abhay)
 
* Tinyint columns can now be considered integers instead of booleans by setting Sequel.convert_tinyint_to_bool = false (samsouder)
 
* Allow the use of URL parameters in connection strings (jeremyevans)
 
* Ignore any previously selected columns when using Dataset#graph for the first time (jeremyevans)
 
* Dataset#graph now accepts a block which is passed to join_table (jeremyevans)
 
* Make Dataset#columns ignore any filtering, ordering, and distinct clauses (jeremyevans)
 
* Use the safer connection-specific string escaping methods for PostgreSQL (jeremyevans)
 
* Database#transaction now yields a connection when using the Postgres adapter, just like it does for other adapters (jeremyevans)
 
* Dataset#count now works for a limited dataset (divoxx)
 
* Database#add_index is now supported in the SQLite adapter (abhay)
 
* Sequel's MySQL adapter should no longer conflict with ActiveRecord's use of MySQL (careo)
 
* Treat Hash as expression instead of column alias when used in DISTINCT, ORDER BY, and GROUP BY clauses (jeremyevans)
 
* PostgreSQL bytea fields are now fully supported (dlee)
 
* For PostgreSQL, don't raise an error when assigning a value to a SERIAL PRIMARY KEY field when inserting records (jeremyevans)
 
=== 2.1.0 (2008-06-17)
 
* Break association add_/remove_/remove_all_ methods into two parts, for easier overriding (jeremyevans)
 
* Add Model.strict_param_setting, on by default, which raises errors if a missing/restricted method is called via new/set/update/etc. (jeremyevans)
 
* Raise errors when using association methods on objects without valid primary keys (jeremyevans)
 
* The model's primary key is a restricted column by default, Add model.unrestrict_primary_key to get the old behavior (jeremyevans)
 
* Add Model.set_(allowed|restricted)_columns, which affect which columns create/new/set/update/etc. modify (jeremyevans)
 
* Calls to Model.def_dataset_method with a block are cached and reapplied to the new dataset if set_dataset is called, even in a subclass (jeremyevans)
 
* The :reciprocal option to associations should now be the symbol name of the reciprocal association, not an instance variable symbol (jeremyevans)
 
* Add Model#associations, which is a hash holding a cache of associated objects, with each association being a separate key (jeremyevans)
  
* Make all associations support a :graph_select option, specifying a column or array of columns to select when using eager_graph (jeremyevans)
 
* Bring back Model#set and Model#update, now the same as Model#set_with_params and Model#update_with_params (jeremyevans)
 
* Allow model datasets to call to_hash without any arguments, which allows easy creation of identity maps (jeremyevans)
 
* Add Model.set_sti_key, for easily setting up single table inheritance (jeremyevans)
 
* Make all associations support a :read_only option, which doesn't add methods that modify the database (jeremyevans)
 
* Make *_to_many associations support a :limit option, for specifying a limit to the resulting records (and possibly an offset) (jeremyevans)
 
* Make association block argument and :eager option affect the _dataset method (jeremyevans)
 
* Add a :one_to_one option to one_to_many associations, which creates a getter and setter similar to many_to_one (a.k.a. has_one) (jeremyevans)
 
* add_ and remove_ one_to_many association methods now raise an error if the passed object cannot be saved, instead of saving without validation (jeremyevans)
 
* Add support for :if option on validations, using a symbol (specifying an instance method) or a proc (dtsato)
 
* Support bitwise operators for NumericExpressions: &, |, ^, ~, <<, >> (jeremyevans)
 
* No longer raise an error for Dataset#filter(true) or Dataset#filter(false) (jeremyevans)
 
* Allow Dataset #filter, #or, #exclude and other methods that call them to use both the block and regular arguments (jeremyevans)
 
* ParseTree support is now officially deprecated, use Sequel.use_parse_tree = false to use the expression (blockless) filters inside blocks (jeremyevans)
 
* Remove :pool_reuse_connections ConnectionPool/Database option, MySQL users need to be careful with nested queries (jeremyevans)
 
* Allow Dataset#graph :select option to take an array of columns to select (jeremyevans)
 
* Allow Dataset#to_hash to be called with only one argument, allowing for easy creation of lookup tables for a single key (jeremyevans)
 
* Allow join_table to accept a block providing the aliases and previous joins, that allows you to specify arbitrary conditions properly qualified (jeremyevans)
 
* Support NATURAL, CROSS, and USING joins in join_table (jeremyevans)
 
* Make sure HAVING comes before ORDER BY, per the SQL standard and at least MySQL, PostgreSQL, and SQLite (juco)
 
* Add cast_numeric and cast_string methods for use in the Sequel DSL, that have default types and wrap the object in the correct class (jeremyevans)
 
* Add Symbol#qualify, for adding a table/schema qualifier to a column/table name (jeremyevans)
 
* Remove Module#metaprivate, since it duplicates the standard Module#private_class_method (jeremyevans)
 
* Support the SQL CASE expression via Array#case and Hash#case (jeremyevans)
 
* Support the SQL EXTRACT function: :date.extract(:year) (jeremyevans)
 
* Convert numeric fields to BigDecimal in PostgreSQL adapter (jeremyevans)
 
* Add :decimal fields to the schema parser (jeremyevans)
 
* The expr argument in join table now allows the same argument as filter, so it can take a string or a blockless filter expression (brushbox, jeremyevans)
 
* No longer assume the expr argument to join_table references the primary key column (jeremyevans)
 
* Rename the Sequel.time_class setting to Sequel.datetime_class (jeremyevans)
 
* Add savepoint/nesting support to postgresql transactions (elven)
 
* Use the specified table alias when joining a dataset, instead of the automatically generated alias (brushbox)
 
=== 2.0.1 (2008-06-04)
 
* Make the choice of Time or DateTime optional for typecasting :datetime types, default to Time (jeremyevans)
 
* Reload database schema for table when calling Model.create_table (jeremyevans)
 
* Have PostgreSQL money type use BigDecimal instead of Float (jeremyevans)
 
* Have the PostgreSQL and MySQL adapters use the Sequel.time_class setting for datetime/timestamp types (jeremyevans)
 
* Add Sequel.time_class and String#to_sequel_time, used for converting time values from the database to either Time (default) or DateTime (jeremyevans)
 
* Make identifier quoting uppercase by default, to work better with the SQL standard, override in PostgreSQL (jeremyevans) (#232)
 
* Add StringExpression#+, for simple SQL string concatenation (:x.sql_string + :y) (jeremyevans)
 
* Make StringMethods.like to a case sensensitive search on MySQL (use ilike for the old behavior) (jeremyevans)
 
* Add StringMethods.ilike, for case insensitive pattern matching (jeremyevans)
 
* Refactor ComplexExpression into three subclasses and a few modules, so operators that don't make sense are not defined for the class (jeremyevans)
 
=== 2.0.0 (2008-06-01)
 
* Comprehensive update of all documentation (jeremyevans)
 
* Remove methods deprecated in 1.5.0 (jeremyevans)
 
* Add typecasting on attribute assignment to Sequel::Model objects, optional but enabled by default (jeremyevans)
 
* Returning false in one of the before_ hooks now causes the appropriate method(s) to immediately return false (jeremyevans)
 
* Add remove_all_* association method for *_to_many associations, which removes the association with all currently associated objects (jeremyevans)
 
* Add Model.lazy_load_schema=, when set to true, it loads the schema on first instantiation (jeremyevans)
 
* Add before_validation and after_validation hooks, called whenever the model is validated (jeremyevans)
 
* Add Model.default_foreign_key, a private class method that allows changing the default foreign key that Sequel will use in associations (jeremyevans)
 
* Cache negative lookup when eagerly loading many_to_one associations (jeremyevans)
 
* Make all associations support the :select option, not just many_to_many (jeremyevans)
 
* Allow the use of blocks when eager loading, and add the :eager_block and :allow_eager association options for configuration (jeremyevans)
  
* Add the :graph_join_type, :graph_conditions, and :graph_join_table_conditions association options, used when eager graphing (jeremyevans)
 
* Add AssociationReflection class (subclass of Hash), to make calling a couple of private Model methods unnecessary (jeremyevans)
 
* Change hook methods so that if a tag/method is specified it overwrites an existing hook block with the same tag/method (jeremyevans)
 
* Refactor String inflection support, you must use String.inflections instead of Inflector.inflections now (jeremyevans)
 
* Allow connection to ODBC-MSSQL via a URL (petersumskas) (#230)
 
* Comprehensive update of all documentation, except for the block filters and adapters (jeremyevans)
 
* Handle Date and DateTime value literalization correctly in adapters (jeremyevans)
 
* Literalize DateTime values the same as Time values (jeremyevans)
 
* MySQL tinyints are now returned as boolean values instead of integers (jeremyevans)
 
* Set additional MySQL charset options required for creating tables and databases (tmm1)
 
* Remove methods deprecated in 1.5.0 (jeremyevans)
 
* Add Module#metaattr_accessor for creating attr_accessors for the metaclass (jeremyevans)
 
* Add SQL string concatenation support to blockless filters, via Array#sql_string_join (jeremyevans)
 
* Add Pagination#last_page? and Pagination#first_page? (apeiros)
 
* Add limited column reflection support, tested on PostgreSQL, MySQL, and SQLite (jeremyevans)
 
* Allow the use of :schema__table___table_alias syntax for tables, similar to the column support (jeremyevans)
 
* Merge metaid gem into core_ext.rb and clean it up, so sequel now has no external dependencies (jeremyevans)
 
* Add Dataset#as, so using a dataset as a column with an alias is not deprecated (jeremyevans)
 
* Add Dataset#invert, which returns a dataset with inverted HAVING and WHERE clauses (jeremyevans)
 
* Add blockless filter syntax support (jeremyevans)
 
* Passing an array to Dataset#order and Dataset#select no longer works, you need to pass multiple arguments (jeremyevans)
 
* You should use '?' instead of '(?)' when using interpolated strings with array arguments (jeremyevans)
 
* Dataset.literal now surrounds the literalization of arrays with parentheses (jeremyevans)
 
* Add echo option (back?) to sequel command line tool, via -E or --echo (jeremyevans)
 
* Allow databases to have multiple loggers (jeremyevans)
 
* The sequel command line tool now also accepts a path to a database config YAML file in addition to a URI (mtodd)
 
* Major update of the postgresql adapter (jdavis, jeremyevans) (#225)
 
* Make returning inside of a database transaction commit the transaction (ahoward, jeremyevans)
 
* Dataset#to_table_reference is now protected, and it has a different API (jeremyevans)
 
* Dataset#join_table and related functions now take an explicit optional table_alias argument, you can no longer include the table alias in the table argument (jeremyevans)
 
* Aliased and/or qualified columns with embedded spaces can now be specified as symbols (jeremyevans)
 
* When identifier quoting is enabled, the SQL standard double quote is used by default (jeremyevans)
 
* When identifier quoting is enabled, quote tables as well as columns (jeremyevans)
 
* Make identifier quoting optional, enabled by default (jeremyevans)
 
* Allow Sequel::Database.connect and related methods to take a block that disconnects the database when the block finishes (jeremyevans)
 
* Add Dataset#unfiltered, for removing filters from dataset (jeremyevans)
 
* Add add_foreign_key and add_primary_key methods to the AlterTableGenerator (jeremyevans)
 
* Allow migration files to have more than 3 digits (jeremyevans)
 
* Add methods directly to Dataset instead of including modules (jeremyevans)
 
* Make some Dataset instance methods private: invert_order, insert_default_values_sql (jeremyevans)
 
* Don't add methods that depend on ParseTree unless you can load ParseTree (jeremyevans)
 
* Don't wipeout the cached columns every time a dataset is cloned, but only on changes to :select, :sql, :from, or :join (jeremyevans)
 
* Fix Oracle Adapter (yasushi.abe)
 
* Fixed sqlite uri so that sqlite:// works just like file:// (2 slashes for a relative path, 3 for an absolute) (dlee)
 
* Raise a Sequel::Error if an invalid limit or offset is used (jeremyevans)
 
* Refactor and beef up Dataset#first and Dataset#last, with some change in functionality (jeremyevans)
 
* Add String#to_datetime, for consistency (jeremyevans)
 
* Fix Range#interval so that it returns 1 less for an exclusive range
 
* Change SQLite adapter so it doesn't swallow exceptions other than SQLite3::Exception (such as Interrupt) (jeremyevans)
 
* Change PostgreSQL and MySQL adapters to raise Sequel::Error instead of database specific errors if a database error occurs (jeremyevans)
 
* Using a memory database with SQLite now defaults to a single connection, so all queries it uses run against the same database (jeremyevans)
 
* Fix attempting to query MySQL using the same connection being used to concurrently execute another query (jeremyevans)
 
* Add options to the connection pool to configure reusing connections and converting exceptions (jeremyevans)
 
* Use the database driver provided string quoting methods for MySQL and SQLite (jeremyevans) (#223)
 
* Add ColumnAll#==, for checking the equality of two ColumnAlls (jeremyevans)
 
* Allow an array of arrays instead of a hash when specifying conditions (jeremyevans)
 
* Add Sequel::DBI::Database#lowercase, for lowercasing column names (jamesearl)
 
* Remove Dataset#extend_with_destroy, which may break code that uses Dataset#set_model directly and expects the destroy method to be added (jeremyevans)
 
* Fix some issues when running on Ruby 1.9 (Zverok, jeremyevans)
 
* Make the DBI adapter work (partially) with PostgreSQL (Seb)
 
=== 1.5.1 (2008-04-30)
 
* Fix Dataset#eager_graph when not all objects have associated objects (jeremyevans)
 
* Have Dataset#graph give a nil value instead of a hash with all nil values if no matching rows exist in the graphed table (jeremyevans)
 
=== 1.5.0 (2008-04-29)
 
* Make the validation errors API compatible with Merb (Inviz)
 
* Add validates_uniqueness_of, for protecting against duplicate entries in the database (neaf, jeremyevans)
 
* Alias Model#dataset= to Model#set_dataset (tmm1)
 
* Make some Model class methods private: def_hook_method, hooks, add_hook, plugin_module, plugin_gem (jeremyevans)
 
* Add the eager! and eager_graph! mutation methods to model datasets (jeremyevans)
 
* Remove Model.database_opened (jeremyevans)
 
* Remove Model.super_dataset (jeremyevans)
 
* Deprecate .create_with_params, .create_with, #set, #update, #update_with, and #new_record from Sequel::Model (jeremyevans)
 
* Add Model.def_dataset_method, for defining methods on the model that reference methods on the dataset (jeremyevans)
 
* Deprecate Model.method_missing, add dataset methods to Model via metaprogramming (jeremyevans)
 
* Remove Model.join, so it is the same as Dataset#join (jeremyevans)
 
* Use reciprocal associations for all types of associations in the getter/setter/add_/remove_ methods (jeremyevans)
 
* Fix many_to_one associations to cache negative lookups (jeremyevans)
 
* Change Model#=== to always be false if the primary key is nil (jeremyevans)
 
* Add Model#hash, which should be unique for a given class and primary key (or values if primary key is nil) (jeremyevans)
 
* Add Model#eql? as a alias to Model#== (jeremyevans)
 
* Make Model#reload clear any cached associations (jeremyevans)
 
* No longer depend on the assistance gem, merge the Inflector and Validations code (jeremyevans)
 
* Add Model#set_with_params, which is Model#update_with_params without the save (jeremyevans)
  
* Fix Model#destroy so that it returns self, not the result of after_destroy (jeremyevans)
 
* Define Model column accessors in set_dataset, so they should always be avaiable, deprecate Model#method_missing (jeremyevans)
 
* Add eager loading of associations via new sequel_core object graphing feature (jeremyevans)
 
* Fix many_to_many associations with classes inside modules without an explicit join table (jeremyevans)
 
* Allow creation of new records that don't have primary keys when the cache is on (jeremyevans) (#213)
 
* Make Model#initialize, Model#set, and Model#update_with_params invulnerable to memory exhaustion (jeremyevans) (#210)
 
* Add Model.str_columns, which gives a list of columns as frozen strings (jeremyevans)
 
* Remove pretty_table.rb from sequel, since it is in sequel_core (jeremyevans)
 
* Set a timeout in the Sqlite adapter, default to 5 seconds (hrvoje.marjanovic) (#218)
 
* Document that calling Sequel::ODBC::Database#execute manually requires you to manually drop the returned object (jeremyevans) (#217)
 
* Paginating an already paginated/limited dataset now raises an error (jeremyevans)
 
* Add support for PostgreSQL partial indexes (dlee)
 
* Added support for arbitrary index types (including spatial indexes) (dlee)
 
* Quote column names in SQL generated for SQLite (tmm1)
 
* Deprecate Object#rollback! (jeremyevans)
 
* Make some Dataset methods private (qualified_column_name, column_list, table_ref, source_list) (jeremyevans)
 
* Deprecate Dataset methods #set_options, #set_row_proc, #remove_row_proc, and #clone_merge (jeremyevans)
 
* Add Symbol#*, a replacement for Symbol#all (jeremyevans)
 
* Deprecate including ColumnMethods in Object, include it in Symbol, String, and Sequel::SQL::Expression (jeremyevans)
 
* Deprecate Symbol#method_missing, and #AS, #DESC, #ASC, #ALL, and #all from ColumnMethods (jeremyevans)
 
* Fix table joining in MySQL (jeremyevans)
 
* Deprecate Sequel.method_missing and Object#Sequel, add real Sequel.adapter methods (jeremyevans)
 
* Move dataset methods applicable only to paginated datasets into Sequel::Dataset::Pagination (jeremyevans)
 
* Make Sequel::Dataset::Sequelizer methods private (jeremyevans)
 
* Deprecate Dataset#method_missing, add real mutation methods (e.g. filter!) (jeremyevans)
 
* Fix connecting to an MSSQL server via ODBC using domain user credentials (jeremyevans) (#216)
 
* No longer depend on the assistance gem, merge in the ConnectionPool and .blank methods (jeremyevans)
 
* No longer depend on ParseTree, RubyInline, or ruby2ruby, but you still need them if you want to use the block filters (jeremyevans)
 
* Fix JDBC adapter by issuing index things start at 1 (pdamer)
 
* Fix connecting to a database via the ADO adapter (now requires options instead of URI) (timuckun, jeremyevans) (#204)
 
* Support storing microseconds in postgres timestamp fields (schnarch...@rootimage.msu.edu) (#215)
 
* Allow joining of multiple datasets, by making the table alias different for each dataset joined (jeremyevans)
 
* SECURITY: Fix backslash escaping of strings (dlee)
 
* Add ability to create a graph of objects from a query, with the result split into corresponding tables (jeremyevans) (#113)
 
* Add attr_accessor for dataset row_proc (jeremyevans)
 
* Don't redefine Dataset#each when adding a transform or row_proc (jeremyevans)
 
* Remove array_keys.rb from sequel_core, it was partially broken (since the arrays came from hashes), and redefined Dataset#each (jeremyevans)
 
* Fix MySQL default values insert (matt.binary) (#196)
 
* Fix ODBC adapter improperly escaping date and timestamp values (leo.borisenko) (#165)
 
* Fix renaming columns on MySQL with type :varchar (jeremyevans) (#206)
 
* Add Sequel::SQL::Function#==, for comparing SQL Functions (jeremyevans) (#209)
 
* Update Informix adapter to work with Ruby/Informix 0.7.0 (gerardo.santana@gmail.com)
 
* Remove sequel_core's knowledge of Sequel::Model (jeremyevans)
 
* Use "\n" instead of $/ (since $/ can be redefined in ways we do not want) (jeremyevans)
 
=== 1.4.0 (2008-04-08) 
 
* Don't mark a column as changed unless the new value is different from the current value (tamas.denes, jeremyevans) (#203).
 
* Switch gem name from "sequel_model" to just "sequel", which required large version bump (jeremyevans).
 
* Add :select option to many_to_many associations, default to selecting only the associated model table and not the join table (jeremyevans) (#208).
 
* Add :reciprocal one_to_many association option, for setting corresponding many_to_one instance variable (jeremyevans).
 
* Add eager loading implementation (jeremyevans).
 
* Change *_to_many associations so that the all associations are considered :cache=>true (jeremyevans).
 
* Fix associa