that #exists? and others can produce invalid SQL: "SELECT DISTINCT DISTINCT"
The combination of a :uniq => true association and the #distinct call
in #construct_limited_ids_condition combine to create invalid SQL, because
we're explicitly selecting DISTINCT, and also sending #distinct on to AREL,
via the relation#distinct_value.
Where #6792 was the forever fix, this is the minimal fix. Instead of
properly indicating the distinctness of the query through #uniq_value alone,
we use a literal select statement and set #uniq_value to always be falsey
See #9777 for details.
Previously pluck is not returning what we wanted to
Added a test also to make sure it's working fine.
This will also fix the build for 1.8.7 as we
were doing some sort on hash.
Thanks @pixeltrix for helping me out.
Thanks @linduxed for pairing with me.
This reverts commit 7240202784, reversing
changes made to e4e2bcce75.
Conflicts:
activerecord/CHANGELOG.md
activerecord/lib/active_record/relation/calculations.rb
activerecord/test/cases/calculations_test.rb
Reason: This caused a regression since it changed the behavior in a
stable release.
Fixes#9777
Confirm a record has not already been destroyed before decrementing counter cache
Conflicts:
activerecord/CHANGELOG.md
Conflicts:
activerecord/CHANGELOG.md
activerecord/lib/active_record/associations/builder/belongs_to.rb
Closes#9806.
As the `through_options` always contained `{:order=>nil}` the second time,
the preloader ran, the association was always reset. This patch only
adds the `:order` to the `through_options` if it is set.
* 3-2-13:
bumping to 3.2.13
fix protocol checking in sanitization [CVE-2013-1857]
JDOM XXE Protection [CVE-2013-1856]
fix incorrect ^$ usage leading to XSS in sanitize_css [CVE-2013-1855]
stop calling to_sym when building arel nodes [CVE-2013-1854]
Merge pull request #9616 from exviva/multiple_select_name_double_square_brackets
bumping to rc2
Revert "Merge pull request #8209 from senny/backport_8176"
Freeze columns only once per Result
Preparing for 3.2.13.rc1 release
Update CHANGELOGs for 3.2.13 release.
Conflicts:
actionmailer/CHANGELOG.md
actionpack/CHANGELOG.md
activemodel/CHANGELOG.md
activeresource/CHANGELOG.md
activesupport/CHANGELOG.md
railties/CHANGELOG.md
This reverts commit 7240202784, reversing
changes made to e4e2bcce75.
Conflicts:
activerecord/CHANGELOG.md
activerecord/lib/active_record/relation/calculations.rb
activerecord/test/cases/calculations_test.rb
* This has been fixed at master via `365b8b6`, but not at 3-2-stable branch.
* @stale_state should be nil when a model isn't saved. via `0f3901e`.
* set @stale_state to nil when reset the target.
The build only failed for ruby-1.8.7 and pg. The problem was that the statement:
```ruby
author = Author.includes(:comments_with_order_and_conditions, :posts).first
```
resulted in Author with ID 2 where on all other rubies / db-engines Author with ID 1
was retunred. Of course this breaks the assertions.
Backport #9252.
Conflicts:
activerecord/CHANGELOG.md
activerecord/test/cases/associations/eager_test.rb
The preloader code on 3-2-stable is not based on relations but on option hashes.
I had to modify the original patch and comparing the option hashes could be more
fuzzy than comparing the relations. All the tests passed though.
Previously, when `time_zone_aware_attributes` were enabled, after
changing a datetime or timestamp attribute and then changing it back
to the original value, `changed_attributes` still tracked the
attribute as changed. This caused `[attribute]_changed?` and
`changed?` methods to return true incorrectly.
Example:
in_time_zone 'Paris' do
order = Order.new
original_time = Time.local(2012, 10, 10)
order.shipped_at = original_time
order.save
order.changed? # => false
# changing value
order.shipped_at = Time.local(2013, 1, 1)
order.changed? # => true
# reverting to original value
order.shipped_at = original_time
order.changed? # => false, used to return true
end
(cherry picked from commit bc982cbcb34129ea2cfe8aa1f8e0b40e444e68db)
Conflicts:
activerecord/CHANGELOG.md
activerecord/lib/active_record/attribute_methods/time_zone_conversion.rb
Backport of #9073Fixes#8898
Issue #6115 has been fixed and tested with
the attribute `:null => false, :default => ""`
However `:null => false` attribute is not necessary to test this issue,
which causes many ORA-01400 errors with Oracle enhanced adapter.
