History.txt

== 1.1.3

- Remove AR version < 3 guard around some caching methods (sugg. invadersmustdie)
- Small bug in arjdbc/discover logic, thanks autotelik.
- Added bigint serial support + some testcases for native type mapping (postgres only)
- mssql: use subscript instead of #first. (Kim Toms)
- #71: fix yield called out of block error
- Silence Rake::DSL warnings for Rake > 0.9

== 1.1.2

- Update version of H2 driver from 1.1.107 to 1.3.153 (Ketan
  Padegaonkar, Jeremy Stephens)
- Fix errors in db:test:clone_structure with PostgreSQL (Andrea Campi)
- Fixing limit for sqlServer2000 if primary key is not named 'id'
  (Luca Simone)
- DB2: define jdbc_columns (fixes table_exists? bug) (Nick Kreucher)
- ACTIVERECORD_JDBC-152 - omitting limit when dumping bytea fields
  (Gregor Schmidt)
- Postgres doesn't support a limit for bytea columns (Alex Tambellini)
- JRUBY-5642: Default to schema public if no schema given for postgres
  (Anthony Juckel)
- Sqlite3 supports float data type so use float (Alex Tambellini)
- GH #21: Now using sqlite3 driver from
  http://www.xerial.org/trac/Xerial/wiki/SQLiteJDBC (thanks Ukabu)
- GH #65: PG: Respect integer sizes (Alex Tambellini)
- GH #59: PG: Properly escape bytea-escaped string
- GH #53: oracle: allow configuration of schema through schema: key
- GH #50: PG: support multiple schema in search_path (Daniel
  Schreiber)
- GH #25: Reload ArJdbc.column_types if number of constants changed
- GH #47: Allow table statistics for indexes to be approximate; speeds
  up Oracle
- GH #67: Change primary_keys to use the same catalog/schema/table
  separation logic as columns_internal (Marcus Brito). This change
  allows set_table_name to specify a custom schema.
- GH #49: mssql: quote table names like column names
- GH #56: mssql: Fix 'select 1' behavior introduced by AR 3.0.7
- GH #55: Make decimal columns with no precision or scale stay
  decimals
- GH #45: Add Arel limit support for Firebird (Systho))
- GH #39: PG: allow negative integer default values
- GH #19: Make a stub Mysql::Error class
- ACTIVERECORD_JDBC-148: mssql: Ensure regex doesn't match 'from' in a
  field name
- GH#31: mssql: Remove extra code breaking mssql w/o limit
- ACTIVERECORD_JDBC-156: mssql: Logic fix for detecting select_count?

== 1.1.1

- Arel 2.0.7 compatibility: fix bugs arising from use of Arel 2.0.7 +
  ArJdbc 1.1.0.
  - Gracefully handle changes to limit in Arel's AST
  - Avoid conflict with Arel 2.0.7's mssql visitor
- Upgrade to PostgreSQL 9.0.801 JDBC drivers (David Kellum)

== 1.1.0 (12/09/10)

- Don't narrow platform to '-java' only: revert back to 0.9.2 where
  ar-jdbc can be installed under any Ruby (for easier Bundler/Warbler
  usage and less confusion on rubygems.org).
- Upgrade MySQL execute code to use RETURN_GENERATED_KEYS.
- Upgrade to MySQL driver version 5.1.13
- Add multi-statement support, idea from oruen. For databases that
  support it, you can now do:
      results = Model.connection.execute("select 1; select 2")
  and receive back an array of multiple result set arrays. For use with
  MySQL, you need to add
      options:
        allowMultiQueries: true
  in database.yml.
- ACTIVERECORD_JDBC-144: Fix limits appearing in schema dump for some
  datatypes (Uwe Kubosch)
- Fixes for DB2 limit/offset
- Fix rake db:create for 'jdbc' adapter (Joeri Samson)
- add create/drop database methods to h2 adapter (sahglie)
- Use connection getDatabaseProductName instead of getClass.getName
  when detecting JNDI dialects (Denis Odorcic)
- ACTIVERECORD_JDBC-146: Fix create_table to not append encoding (Marc Slemko)
- All green on SQLite3 Rails master ActiveRecord tests
- ACTIVERECORD_JDBC-140: Sync postgres add/change column code from Rails master
- ACTIVERECORD_JDBC-139: TEXT/DATE on PostgreSQL should not have limits

== 1.0.3 (11/29/10)

- ACTIVERECORD_JDBC-143: Implement table_exists? fixing association
  table names with schema prefixes
- Cleanup of column code for hsqldb (Denis Odorcic)
- Rails 3.0.3 support - add Arel 2 visitors for all adapters
- Fix MySQL date types to not have limits (Chris Lowder)
- ACTIVERECORD_JDBC-141: Better schema support in H2

== 1.0.2

- ACTIVERECORD_JDBC-134: Fix conflicting adapter/column superclasses
- ACTIVERECORD_JDBC-135: Fix regression on PG with boolean and :limit
- Slew of Derby fixes courtesy of Denis Odorcic

== 1.0.1

- Fix db:test:purge issue affecting all adapters in 1.0.0 due to
  incorrect fix to JRUBY-5081 in 8b4b9c5

== 1.0.0

- Thanks to David Kellum, Dmitry Denisov, Dwayne Litzenberger, Gregor
  Schmidt, James Walker, John Duff, Joshua Suggs, Nicholas J Kreucher,
  Peter Donald, Geoff Longman, Uwe Kubosch, Youhei Kondou, Michael
  Pitman, Alex B, and Ryan Bell for their contributions to this
  release.
- BIG set of DB2 updates (Thanks Nick Kreucher)
- Deprecate jdbc_adapter/rake_tasks
- (1.0.0.beta1)
- Make database-specific extensions only load when necessary
- Allow for discovery of database extensions outside of ar-jdbc
  proper. This should allow for custom database development to be
  developed and released without relying on AR-JDBC core.
- Get AR's own tests running as close to 100% as possible. MySQL is
  currently 100%, SQLite3 is close.
- JRUBY-4876: Bump up Derby's max index name length (Uwe Kubosch)
- (1.0.0.beta2)
- 98 commits since beta1
- MSSQL updates from dlitz and realityforge
- ACTIVERECORD_JDBC-131: Fix string slug issue for DB2 (Youhei Kondou)
- JRUBY-1642: Don't use H2 INFORMATION_SCHEMA in table or column
  searches
- JRUBY-4972: Attempt to deal with type(0)/:limit => 0 by not setting
  it808e213
- JRUBY-5040: Fix issue with limits on timestamps in MySQL
- JRUBY-3555: Allow setting Derby schema with 'schema:' option
- ACTIVERECORD_JDBC-98: Make sure we actuall raise an error when
  inappropriately configured
- ACTIVERECORD_JDBC-112: Add schema dumper tests for already-fixed
  MySQL type limits
- ACTIVERECORD_JDBC-113: Fix PG float precision issue
- ACTIVERECORD_JDBC-103: Fix decimal options for PG add/change column
  (Michael Pitman)
- ACTIVERECORD_JDBC-127: Fix quoting of Date vs. Time(stamp) for
  Oracle (Lenny Marks)
- Oracle: Sort out the NUMBER vs NUMBER(x) vs NUMBER(x,y) situation.
- JRUBY-3051: Think we finally got the PG mixed-case patches applied.
- JRUBY-5081: Consolidate code for dropping DB via postgres
- ACTIVERECORD_JDBC-101: Add override of LONGVARCHAR => CLOB for
  informix
- ACTIVERECORD_JDBC-107: Fix MySQL update_all issue on AR 2.3
- ACTIVERECORD_JDBC-124: Filter out special _row_num column
- ACTIVERECORD_JDBC-126: Fix sql 2000 limit/offset per Michael Pitman
- ACTIVERECORD_JDBC-125: Add tweak to limit/offset code for HABTM
  queries (alex b)
- ACTIVERECORD_JDBC-129: Don't have limits for text, binary or bit
  fields
- (1.0.0 final)
- Fix a few more SQLite3 AR tests
- SQLite3: handle ":memory:" database
- Release new SQLite3 driver 3.6.14.2 and new Derby driver 10.6.2.1

== 0.9.7

- JRUBY-4781: Fix multiple database connection collision issue w/
  Oracle
- ACTIVERECORD_JDBC-115: Support SAVEPOINTS for MySQL and PG so that
  nested transactions can be faked
- ACTIVERECORD_JDBC-116: Handle schema.table better for MySQL (thanks
  Dilshod Mukhtarov)
- Fix 'Wrong # of arguments (2 for 1)' issue with #create_database for
  MySQL and AR 3.0
- SQLServer 2000 support (thanks Jay McGaffigan)

== 0.9.6

- The Oracle release!
- Oracle should be working much better with this release. Also updated
  to work with Rails 3.
- Get all unit tests running cleanly on Oracle, fixing previous
  datetime/timezone issues.
- ACTIVERECORD_JDBC-83: Add :sequence_start_value option to
  create_table, following oracle_enhanced adapter
- ACTIVERECORD_JDBC-33: Don't double-quote table names in oracle
- ACTIVERECORD_JDBC-17: Fix Oracle primary keys so that /^NUMBER$/ => :integer
- Fix remaining blockers ACTIVERECORD_JDBC-82, JRUBY-3675,
  ACTIVERECORD_JDBC-22, ACTIVERECORD_JDBC-27, JRUBY-4759

== 0.9.5

- The MSSQL release, courtesy of Mike Williams and Lonely
  Planet.
- JRuby + AR-JDBC is now seen as the hassle-free way of using Rails
  with SQLServer!
- Many fixes for MSSQL, including ACTIVERECORD_JDBC-18,
  ACTIVERECORD_JDBC-41, ACTIVERECORD_JDBC-56, ACTIVERECORD_JDBC-94,
  ACTIVERECORD_JDBC-99, JRUBY-3805, JRUBY-3793, JRUBY-4221
- All tests pass on Rails 3.0.0.beta3!

== 0.9.4

- ACTIVERECORD_JDBC-96: DB2 JdbcSpec cannot dump schema correctly
  (Youhei Kondou)
- ACTIVERECORD_JDBC-97: Dont use Rails 3 deprecated constants (David
  Calavera)
- Updates for rake db:schema:dump compatibility with Rails 2.3+ and
  MySQL (Joakim Kolsjö)
- Rails 3.0.0.beta2 compatibility
- Return of Derby, H2, Hsqldb support (requires AR >= 3.0.0.beta2)

== 0.9.3

- Rails 3 compatibility
- PLEASE NOTE: ActiveRecord in Rails 3 has changed in a way that
  doesn't allow non-standard DBs (such as the Derby and H2 embedded
  DBs) to work. We're investigating the effort required to support
  these databases and hope to have something for a future release.
- ACTIVERECORD_JDBC-91: Fix schema search path for PostgreSQL (Alex
  Kuebo)
- ACTIVERECORD_JDBC-87: DB2 ID insert fix (Youhei Kondou)
- ACTIVERECORD_JDBC-90: MSSQL fix for DATEs (jlangenauer)
- ACTIVERECORD_JDBC-93: Fix string IDs for sqlite3, hsql/h2 (moser)
- ACTIVERECORD_JDBC-86: Fix Derby queries starting with VALUES (Dwayne Litzenberger)
- ACTIVERECORD_JDBC-95: Fix INSERT ... RETURNING for PostgreSQL

== 0.9.2

- The main, highly awaited fix for this release is a solution to the
  rake db:create/db:drop issue. The main change is a new 'jdbc' rails
  generator that should be run once to prepare a Rails application to
  use JDBC. The upside of this generator is that you no longer will
  need to alter database.yml for JDBC. See the README.txt for details.
- Cleanup and reconnect if errors occur during begin/rollback
  (Jean-Dominique Morani, Christian Seiler)
- ACTIVERECORD_JDBC-1: Add #drop_database method for oracle (does the
  same thing as recreate_database)
- Sqlite3 and MSSQL fixes (Jean-Dominique Morani)
- JRUBY-3512: Treat LONGVARCHAR as a CLOB for Mssql
- JRUBY-3624: Upgrade Derby to 10.5.3.0 and add native limit/offset
  support (Christopher Saunders)
- JRUBY-3616: Fix postgres non-sequence primary keys (David Kellum)
- JRUBY-3669: Fix Oracle case with unconfigured schema (Dan Powell)
- Fixed quote_column_name of jdbc_oracle to accept numbers (Marcelo
  Murad)
- Fix for mysql tables with non standard primary keys such that the
  schema dump is correct (Nick Zalabak)
- MSSQL fixes from Mike Luu:
  - add support for MSSQL uniqueidentifier datatype
  - always quote strings using unicode identifier for MSSQL
- Changes primary_key generation to use always instead of by default
  for DB2 (Amos King)
- Improves the SQLite adapter by fixing rename_column, change_column,
  change_column_default, changing remove_column, and adding
  remove_columns (Ryan Baumann)
- More oracle love courtesy Ben Browning and Jens Himmelreich
- JRUBY-3608: Add missing change_column_null method for postgres
- JRUBY-3508: Fix quoting of integer and float columns

== 0.9.1

- We did a lot of internal cleanup this release in the hopes of
  simplifying the code and increasing performance.
- Many SQLite updates (thanks Nils Christian Haugen)
- JRUBY-2912: Fix MSSQL create/drop database (Joern Hartmann)
- JRUBY-2767: Mistake in selecting identity with H2/HSQLDB
- JRUBY-2884: jdbc_postgre.rb issue handling nil booleans (also a fix
  for hsqldb/h2) + tests
- JRUBY-2995: activerecord jdbc derby adapter should quote columns
  called 'year'
- JRUBY-2897:  jdbc_postgre.rb needs microsecond support
- JRUBY-3282: Upgrade to derby 10.4.2.0 to allow unique constraints
  with nullable columns
- Update h2 from 1.0.63 to 1.1.107 in driver
- JRUBY-3026: [Derby] Allow select/delete/update conditions with
  comparison to NULL using '='
- JRUBY-2996: ...(actually this fixes only remaining issue of this bug
  which was symbols making into quote were exploding
- JRUBY-2691: Update sybase driver to pass simple unit tests with jtds
  and verify it works with the new dialect keyword. patch by Leigh
  Kennedy
- Make :float type work on h2,hsql [returned as string]. Make :float
  work on hsqldb (no paren value supported). Make REAL_TYPE just
  return RubyFloat
- JRUBY-3222: Upgrade #type_to_sql to variation of AR 2.1.2 version
- Add patch supplied in JRUBY-3489 (patch by Jean-Dominique Morani)
- Various Oracle fixes by edsono
- JRUBY-2688: Don't hard-code MySQL connection character encoding to
  utf8

== 0.9

- Now updated to support ActiveRecord 2.2. JNDI-based connections will
  automatically connect/disconnect for every AR connection pool
  checkout/checkin. For best results, set your pool: parameter >= the
  actual maximum size of the JNDI connection pool. (We'll look at how
  to eliminate the need to configure AR's pool in the future.)
- NEW! Informix support courtesy of Javier Fernandez-Ivern.
- Backport another Oracle CLOB issue, thanks Edson César.
- Rubyforge #22018: chomp final trailing semicolon for oracle
- JRUBY-2848: Fix NPE error in set_native_database_types
- Rework oracle lob saving callback to be Rails 2.1 friendly (assist
  from court3nay)
- JRUBY-2715: Add create/drop database methods to Postgres (Peter Williams)
- JRUBY-3183: Fix structure dump for Postgres (Ryan Bell)
- JRUBY-3184: recreate_database for test database working for PG (Ryan Bell)
- JRUBY-3186: disable referential integrity for PG (Ryan Bell)
- Authoritative repository now hosted at
  git://github.com/nicksieger/activerecord-jdbc-adapter.git; rubyforge
  svn trunk cleaned out.

== 0.8.2

- Added an optional config key called :dialect. Using :dialect allows you to
  override the default SQL dialect for the driver class being used. There are
  a few cases for this:
  - Using using Sybase w/ the jTDS driver.
  - Using rebranded drivers.
  - It makes more sense to use :dialect, rather then :driver when using JNDI.
- JRUBY-2619: Typo with :test config causing problems with dev database (Igor Minar)
- 20524, JRUBY-2612: Since when did I think that there was a #true? method on Object?

== 0.8.1

- Now sporting a JDBC sqlite3 adapter! Thanks Joseph Athman.
- Added support for InterSystems Cache database (Ryan Bell)
- Fix for JRUBY-2256
- JRUBY-1638, JRUBY-2404, JRUBY-2463: schema.table handling and Oracle NUMBER fixes (Darcy Schultz & Jesse Hu)
- Add structure dump and other DDL-ish for DB2 (courtesy abedra and stuarthalloway)
- Fix missing quote_table_name function under Rails 1.2.6 and earlier
- Small tweaks to jdbc.rake to select proper config
- JRUBY-2011: Fix MSSQL string un-quoting issue (Silvio Fonseca)
- JRUBY-1977, 17427: Fix information_schema select issue with MSSQL (Matt Burke)
- 20479: Improve get_table_name for MSSQL (Aslak Hellesøy)
- 20243: numerics improvements for MSSQL (Aslak Hellesøy)
- 20172: don't quote table names for MSSQL (Thor Marius Henrichsen)
- 19729: check for primary key existence in postgres during insert (Martin Luder)
- JRUBY-2297, 18846: retrying failing SQL statements is harmful when not autocommitting (Craig McMillan)
- 10021: very preliminary sybase support. (Mark Atkinson) Not usable until collision w/ sqlserver driver is resolved.
- JRUBY-2312, JRUBY-2319, JRUBY-2322: Oracle timestamping issues (Jesse Hu & Michael König)
- JRUBY-2422: Fix MySQL referential integrity and rollback issues
- JRUBY-2382: mysql string quoting fails with ArrayIndexOutofBoundsException

== 0.8

- NOTE: This release is only compatible with JRuby 1.1RC3 or later.
- Because of recent API changes in trunk in preparation for JRuby 1.1, this release is not
  backward compatible with previous JRuby releases. Hence the version bump.
- Internal: convert Java methods to be defined with annotations
- Fix problem with reserved words coming back pre-quoted from #indexes in postgres
- JRUBY-2205: Fix N^2 allocation of bytelists for mysql quoting (taw)
- Attempt a fix for Rubyforge 18059
- Upgrade derby to 10.3.2.1
- Fix db:create etc. in the case where JDBC is loaded in Rails' preinitializer.rb
- Fix db:drop to actually work
- Fix for Rubyforge #11567 (Matt Williams)

== 0.7.2

- JRUBY-1905: add_column for derby, hsqldb, and postgresql (Stephen Bannasch)
- Fix db:create for JDBC
- Support Rails 2 with the old "require 'jdbc_adapter'" approach
- JRUBY-1966: Instead of searching for just tables, search for views and tables.
- JRUBY-1583: DB2 numeric quoting (Ryan Shillington)
- JRUBY-1634: Oracle DATE type mapping (Daniel Wintschel)
- JRUBY-1543: rename_column issue with more recent MySQL drivers (Oliver Schmelzle)
- Rubyforge #15074: ConnectionAdapters::JdbcAdapter.indexes is missing name and
  schema_name parameters in the method signature (Igor Minar)
- Rubyforge #13558: definition for the indexes method (T Meyarivan)
- JRUBY-2051: handle schemaname and tablename more correctly for columns
- JRUBY-2102: Postgres Adapter cannot handle datetime type (Rainer Hahnekamp)
- JRUBY-2018: Oracle behind ActiveRecord-JDBC fails with "Invalid column index" (K Venkatasubramaniyan)
- JRUBY-2012: jdbc_mysql structure dump fails for mysql views (Tyler Jennings)

== 0.7.1

- Add adapter and driver for H2 courtesy of Caleb Land
- Fix "undefined method `last' for {}:Hash" error introduced with new Rake 0.8.1 (JRUBY-1859)

== 0.7

- PLEASE NOTE: This release is not compatible with JRuby releases earlier than
  1.0.3 or 1.1b2. If you must use JRuby 1.0.2 or earlier, please install the
  0.6 release.
- Release coincides with JRuby 1.0.3 and JRuby 1.1b2 releases
- Simultaneous support for JRuby trunk and 1.0 branch
- Get rid of log_no_bench method, so we time SQL execution again.
- Implement #select_rows
- MySQL migration and quoting updates

== 0.6

- Gem is renamed to "activerecord-jdbc-adapter" to follow new conventions
  introduced in Rails 2.0 for third-party adapters. Rails 2.0 compatibility is
  introduced.
- Add dependency on ActiveRecord >= 1.14 (from the Rails 1.1.x release)
- New drivers (jdbc-XXX) and adapter (activerecord-jdbcXXX-adapter) gems
  available separately. See the README.txt file for details.
- Plain "jdbc" driver is still available if you want to use the full
  driver/url way of specifying the driver.
- More bugfixes to Oracle and SQLServer courtesy of Ola & ThoughtWorks

== 0.5

- Release coincides with JRuby 1.0.1 release
- It is no longer necessary to specify :driver and :url configuration
  parameters for the mysql, postgresql, oracle, derby, hsqldb, and h2
  adapters. The previous configuration is still valid and compatible, but for
  new applications, this makes it possible to use the exact same database.yml
  configuration as Rails applications running under native Ruby.
- JDBC drivers can now be dynamically loaded by Ruby code, without being on
  the classpath prior to launching JRuby. Simply use "require
  'jdbc-driver.jar'" in JRuby code to add it to the runtime classpath.
- Updates to HSQL, MS SQLServer, Postgres, Oracle and Derby adapters

== 0.4

- Release coincides with JRuby 1.0 release
- Shoring up PostgreSQL (courtesy Dudley Flanders) and HSQL (courtesy Matthew
  Williams)
- Fix timestamps on Oracle to use DATE (as everything else)
- Derby fixes: Fix for open result set issue, better structure dump, quoting,
  column type changing
- Sybase type recognition fix (courtesy Dean Mao)

== 0.3.1

- Derby critical fixes shortly after 0.3

== 0.3

- Release coincides with JRuby 1.0.0RC1 release
- Improvements for Derby, Postgres, and Oracle, all of which are running 
  > 95% of AR tests

== 0.2.4

- Release coincides with JRuby 0.9.9 release
- JRuby 0.9.9 is required
- MySQL close to 100% working
- Derby improvements
- DECIMAL/NUMERIC/FLOAT/REAL bugs fixed with type recognition for Oracle,
  Postgres, etc.
- HSQLDB has regressed this release and may not be functioning; we'll get it
  fixed for the next one

== 0.2.3

- Release coincides (and compatible) with JRuby 0.9.8 release
- 8 bugs fixed: see http://rubyurl.com/0Da
- Improvements and compatibility fixes for Rails 1.2.x

== 0.2.1, 0.2.2

- Early releases, added better support for multiple databases

== 0.0.1

- Initial, very alpha release