Commit Graph

2344 Commits

Author SHA1 Message Date
Tarmo Tänav
fa795ccfad Include mysql older than 5.1.23 in the 5.1 series in the list of those that can't handle NULL defaults
Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
2008-08-26 00:53:19 -07:00
Jeremy Kemper
52e15abbed um.. yeah 2008-08-26 00:10:16 -07:00
Jeremy Kemper
6d66ddaa34 typo 2008-08-26 00:02:30 -07:00
Jeremy Kemper
842d55cb16 fix another ordering failure 2008-08-26 00:02:22 -07:00
Jeremy Kemper
ca48da6300 fix tests relying on implicit ordering 2008-08-25 23:53:31 -07:00
Josh Susser
1092c181b5 add dynamic finder bang version to raise RecordNotFound
[#905 state:resolved]

Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
2008-08-25 23:33:16 -07:00
Josh Susser
143f5fbb21 refactor dynamic finder name matching into its own class
Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
2008-08-25 23:32:20 -07:00
Jeremy Kemper
3beed9cdb7 ensure tests load sibling Active Support instead of a gem 2008-08-25 23:32:03 -07:00
Tarmo Tänav
77b003fb61 Use DECIMAL instead of INTEGER when casting as mysql doesn't work with just "INTEGER" and other databases don't like "UNSIGNED" which mysql requires
And don't mask exceptions.

Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
2008-08-25 23:23:34 -07:00
Tarmo Tänav
3d2ac918b9 Cache migrated versions list in Migrator and use it to fetch the latest migrated version name [#845 state:resolved]
Also optimized Migrator#current_version class method to fetch
only the latest version number and not all of them.

With this change no matter how many migrations there are the
schema_migrations table is only SELECTed from once.

Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
2008-08-25 22:03:47 -07:00
Tarmo Tänav
a445cdd884 Load the first and not the last has_one result when doing join-based eager loading
This matters when the has_one is defined with an order in which case
there is an expectation that the first one will be loaded.

[#904 state:resolved]

Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
2008-08-25 21:23:35 -07:00
Frederick Cheung
2dbda11945 Implement old-skool eagerloading for has_one :through
Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
2008-08-25 21:23:15 -07:00
Jeremy Kemper
172606e21f Harmonize framework require strategy. Don't add self to load path since Rails initializer and RubyGems handle it. 2008-08-25 21:22:34 -07:00
Frederick Cheung
b7a37b742c Fix preloading of has_one through associations
[#903 state:resolved]

Signed-off-by: Jeremy Kemper <jeremy@bitsweat.net>
2008-08-25 20:49:55 -07:00
Jeremy Kemper
e5cad34916 strip trailing whitespace 2008-08-25 18:16:58 -07:00
Jeremy Kemper
1c54ca4f75 Ruby 1.9 compat: fix test error masked by old String#each behavior 2008-08-25 18:16:21 -07:00
Sven Fuchs
49859b0bb1 I18n: fix activerecord i18n test for classy backend 2008-08-25 11:48:03 +02:00
Jeremy Kemper
950ea33242 Fix test to assign the module instead of a new instance 2008-08-24 11:08:49 -07:00
Michael Koziarski
33ed19f428 Merge branch 'patches' 2008-08-24 14:34:47 +02:00
Xavier Noria
e02f0dcc24 Rollback the transaction when a before_* callback returns false.
Previously this would have committed the transaction but not carried out save or destroy operation.

[#891 state:committed]
Signed-off-by: Michael Koziarski <michael@koziarski.com>
2008-08-24 14:34:24 +02:00
Jeremy Kemper
c471f13db6 Ruby 1.9 compat: Hash is now flattenable, so explicitly exclude it 2008-08-23 21:26:14 -07:00
Tarmo Tänav
74c3c701f7 Don't set "NULL" as a constraint on nullable columns [#398 state:resolved]
This is already the default and adding it breaks SQL standards compatibility.
2008-08-23 19:51:09 +03:00
Tarmo Tänav
cf28109158 Always require activesupport, even if its constant already exists
This is needed because the existance of the ActiveSupport
constant by itself does not guarantee that the whole library
has been loaded.

Also load the StringInquirer in the Rails#env method as
the it might be called inside the initializer block
before activesupport itself has been loaded.
2008-08-23 18:05:52 +03:00
Josh Susser
e48e77e022 coerce blank strings to nil values for boolean and integer fields
[#860 state:resolved]
2008-08-22 16:26:40 -07:00
Tarmo Tänav
707ee0e269 Made migrations transactional for PostgreSQL [#834 state:resolved]
Patch originally from http://dev.rubyonrails.org/ticket/5470
2008-08-22 14:46:25 -07:00
Jeremy Kemper
9dac5547ad Merge branch 'master' into i18n 2008-08-22 13:01:53 -07:00
Jeremy Kemper
a6e05b18d6 Merge commit 'sven/i18n' into i18n 2008-08-22 12:44:14 -07:00
Patrick Reagan
683ff235e6 Ensure t.timestamps respects options. [#828 state:resolved]
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
2008-08-22 12:48:00 +01:00
Tarmo Tänav
52ac9d0444 Fixed ordering in test_find_in_association_with_custom_finder_sql_and_multiple_interpolations 2008-08-22 10:39:45 +03:00
Tarmo Tänav
d3b894563a Properly quote CREATE DATABASE parameters in postgresql [#771 state:resolved] 2008-08-22 10:37:09 +03:00
Jeremy Kemper
6e3d2a7996 Revert "Performance: freeze cached rows instead of duping"
This reverts commit cd8e653d5b.
2008-08-21 21:40:49 -07:00
Jeremy Kemper
a5eb297424 Revert "coerce blank strings to nil values for boolean and integer fields"
This reverts commit aee14630d4.

[#860 state:incomplete]
2008-08-21 21:34:37 -07:00
S. Brent Faulkner
bbedb6a624 remember created records and select a random one instead of relying on sequential id values starting at 1
Signed-off-by: Michael Koziarski <michael@koziarski.com>
2008-08-21 19:33:26 +02:00
Josh Susser
aee14630d4 coerce blank strings to nil values for boolean and integer fields
Signed-off-by: Michael Koziarski <michael@koziarski.com>
2008-08-21 19:33:25 +02:00
Sven Fuchs
aad429a46e Merge branch 'master' into i18n 2008-08-21 19:04:39 +02:00
Iain Hecker
cf6840773b Custom error messages scope improved 2008-08-21 18:40:44 +02:00
Peter Wagenet
8622787f87 Don't interpret decimals as table names in ActiveRecord::Associations::ClassMethods#references_eager_loaded_tables? [#532 state:resolved]
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
2008-08-21 17:21:17 +01:00
Tom Lea
3724dafe71 Fix incorrect signature for NamedScope#respond_to? [#852 state:resolved]
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
2008-08-21 16:39:27 +01:00
Miles Georgi
0d74e72e6d Fix postgres bug when change_column is called with invalid parameters. [#861 state:resolved]
Signed-off-by: Tarmo Tänav <tarmo@itech.ee>
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
2008-08-21 16:22:14 +01:00
Philip Hallstrom
49c0e1e594 Fix generated WHERE IN query for named scopes. [#583 state:resolved]
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
2008-08-21 16:11:28 +01:00
Xavier Noria
a970f916fb Fix has_many#count_records. [#865 state:resolved]
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
2008-08-21 15:48:04 +01:00
Jakub Kuźma
ea40f71431 Fix that has_one natural assignment to already associated record. [#854 state:resolved]
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
2008-08-21 15:21:21 +01:00
Ryan Bates
2415652660 Support find_all on named scopes. [#730 state:resolved]
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
2008-08-21 12:37:19 +01:00
Iain Hecker
72366398b2 Removed en-US ruby locale in favor of yaml 2008-08-20 18:01:13 +02:00
Iain Hecker
febe2ea977 Locale file changed to yaml 2008-08-20 18:01:12 +02:00
Sven Fuchs
a6a62004c9 add human_name and value to ar validation #generate_message 2008-08-20 17:51:42 +02:00
Iain Hecker
e2b191681e Added :value as interpolation variable available to error messages 2008-08-20 17:41:43 +02:00
Iain Hecker
c531248938 Introduced AR::Base.human_name to validations 2008-08-20 17:39:43 +02:00
Iain Hecker
ae8a35d8f6 Added Base.human_name method 2008-08-20 17:39:43 +02:00
Joshua Peek
c1a8690d58 Consistently use the framework's configured logger and avoid reverting to RAILS_DEFAULT_LOGGER unless necessary. 2008-08-18 23:36:13 -05:00