aboutsummaryrefslogtreecommitdiffstats
path: root/activerecord
Commit message (Collapse)AuthorAgeFilesLines
* Verify that there is no unwanted implicit readonly set on ↵oleg dashevskii2010-12-231-1/+13
| | | | Model.has_many_through.find(id) [#5442 state:resolved]
* Raise an error for associations which try to go :through a polymorphic ↵Jon Leighton2010-12-234-6/+22
| | | | association [#6212 state:resolved]
* If a has_many goes :through a belongs_to, and the foreign key of the ↵Jon Leighton2010-12-2314-35/+124
| | | | belongs_to changes, then the has_many should be considered stale.
* Improved strategy for updating a belongs_to association when the foreign key ↵Jon Leighton2010-12-237-45/+37
| | | | changes. Rather than resetting each affected association when the foreign key changes, we should lazily check for 'staleness' (where fk does not match target id) when the association is accessed.
* Test to verify that #2189 (count with has_many :through and a named_scope) ↵Jon Leighton2010-12-232-0/+7
| | | | is fixed
* Revert "Optimize <association>_ids for hm:t with belongs_to source". The ↵Jon Leighton2010-12-232-14/+3
| | | | | | | | | | optimisation has too many edge cases, such as when the reflection, source reflection, or through reflection has conditions, orders, etc. [#6153 state:resolved] This reverts commit 373b053dc8b99dac1abc3879a17a2bf8c30302b5. Conflicts: activerecord/lib/active_record/associations.rb
* Test demonstrating problem with foo.association_ids where it's a has_many ↵Jon Leighton2010-12-231-0/+4
| | | | :through with :conditions, with a belongs_to as the source reflection
* When a has_many association is not :uniq, appending the same record multiple ↵Jon Leighton2010-12-233-3/+24
| | | | times should append it to the @target multiple times [#5964 state:resolved]
* Fix behaviour of foo.has_many_through_association.select('custom select') ↵Michał Łomnicki2010-12-232-2/+7
| | | | [#6089 state:resolved]
* Fix creation of has_many through records with custom primary_key option on ↵Szymon Nowak2010-12-235-2/+38
| | | | belongs_to [#2990 state:resolved]
* Verify that has_one :through preload respects the :conditions [#2976 ↵Will Bryant2010-12-232-0/+13
| | | | state:resolved]
* Don't allow a has_one association to go :through a collection association ↵Jon Leighton2010-12-235-9/+26
| | | | [#2976 state:resolved]
* avoid duping and new objectsAaron Patterson2010-12-231-2/+3
|
* fixed retrieval of primary key value in Ralation#insert methodRaimonds Simanovskis2010-12-231-1/+1
| | | | previously primary key value was always assigned nil which caused Oracle enhanced adapter failing tests
* Fix for default_scope tests to ensure comparing of equally sorted listsRaimonds Simanovskis2010-12-232-2/+2
| | | | | | This is additional fix for commit ebc47465a5865ab91dc7d058d2d8a0cc961510d7 Respect the default_scope on a join model when reading a through association which otherwise was failing on Oracle (as it returned fixture comments in different order).
* arel can escape the id, so avoid using the database connectionAaron Patterson2010-12-221-1/+1
|
* Arel::Table#[] always returns an attribute, so no need for ||Aaron Patterson2010-12-221-1/+1
|
* probably should use the some_ids variable here. o_OAaron Patterson2010-12-221-1/+1
|
* removing SQL interpolation, please use scoping and attribute conditionals as ↵Aaron Patterson2010-12-224-13/+4
| | | | a replacement
* use arel to compile SQL rather than build stringsAaron Patterson2010-12-221-1/+8
|
* to_sym stuff before passing it to arelAaron Patterson2010-12-222-2/+2
|
* build an AST rather than build SQL stringsAaron Patterson2010-12-221-1/+10
|
* returning id (for some yet to be discovered reason)Aaron Patterson2010-12-201-1/+1
|
* define_attr_method must serialize nil correctlyAaron Patterson2010-12-204-2/+20
|
* if there is no base name, we cannot determine a primary keyAaron Patterson2010-12-202-2/+6
|
* remove some lasignsAaron Patterson2010-12-201-6/+6
|
* no use for set, no need to to_ary, reduce extra objectsAaron Patterson2010-12-201-3/+2
|
* use array arithmetic rather than create setsAaron Patterson2010-12-201-4/+2
|
* Avoid Symbol#to_proc for performance reasons in Ruby 1.8Jon Leighton2010-12-201-3/+3
|
* has_many associations with :dependent => :delete_all should update the ↵Jon Leighton2010-12-204-1/+18
| | | | counter cache when deleting records
* Refactor delete_records in HasManyAssociationJon Leighton2010-12-201-7/+8
|
* Fix problem where wrong keys are used in JoinAssociation when an association ↵Jon Leighton2010-12-205-9/+29
| | | | goes :through a belongs_to [#2801 state:resolved]
* Remove unnecessary overloaded methods create, create! and create_record from ↵Jon Leighton2010-12-201-18/+0
| | | | HasAndBelongsToManyAssociation
* Refactor create_record and build_record in AssociationCollectionJon Leighton2010-12-201-17/+5
|
* Delete create, create! and create_record from HasManyThroughAssociation in ↵Jon Leighton2010-12-202-30/+14
| | | | exchange for more generic versions in AssociationCollection
* Specify insert_record with NotImplementedError in AssociationCollection, to ↵Jon Leighton2010-12-204-11/+14
| | | | indicate that subclasses should implement it. Also add save_record to reduce duplication.
* Merge branch 'master' of git://github.com/lifo/docrailsXavier Noria2010-12-201-2/+2
|\
| * changes in examples - reflect new mailer api and mysql2 adapterVijay Dev2010-12-181-2/+2
| |
| * test for eager load of has_one association with condition on the through tablePivotal Labs2010-12-151-1/+9
| |
| * make sure that join nodes are uniqAaron Patterson2010-12-151-1/+1
| |
| * Verify that creating a has_many through record where there is a ↵Jon Leighton2010-12-154-1/+22
| | | | | | | | default_scope on the join model works correctly (creates the join record with the default scope applied)
| * Respect the default_scope on a join model when reading a through associationJon Leighton2010-12-158-4/+34
| |
| * Fix hm:t to self table aliasing in construct_scopeErnie Miller2010-12-143-3/+12
| |
| * add test which fails for has_many through self join [#4361 state:open]Joe Hannon2010-12-141-0/+7
| |
| * class names are already strings, so we do not need to call to_s on the ↵Aaron Patterson2010-12-141-2/+2
| | | | | | | | strings that are already strings
| * build SQL AST nodes rather than generate stringsAaron Patterson2010-12-141-10/+8
| |
| * construct_joins no longer needs an argumentAaron Patterson2010-12-141-1/+1
| |
| * supporting arel AST nodes when building join statementsAaron Patterson2010-12-142-18/+32
| |
| * just copy the joins to the listAaron Patterson2010-12-141-1/+1
| |
| * bucketing based on join typeAaron Patterson2010-12-141-8/+17
| |