Such request can happen on Internet Explorer. When we redirect
after multipart form submission, the request type is changed
to GET, but Content-Type is preserved as multipart. GET request
cannot have multipart body and that caused Rails to fail.

It's similar fix to Rack's one:
https://github.com/chneukirchen/rack/blob/8025a4ae9477d1e6231344c2b7d795aa9b3717b6/lib/rack/request.rb#L224

Arel::Nodes::In inherits from Arel::Nodes::Equality, so the case
statement was always using the Equality operator for both scenarios,
resulting in a not equal query instead.

Update #where rdoc to match 6ba0f975 [ci skip]

This commit updates the rdoc of AR#where to match the changes applied
in https://github.com/rails/rails/commit/6ba0f97 that is:

* `where(nil)` has the same effect of `where('')`: a no op
* `where` (no args) has the same effect of `where(:chain)`: to create a WhereChain

This reverts commit c59734f7.

This test does not belong to has many associations test.

Consider this scenario:

    if params[:foo]
      conditions = { foo: true }
    end

    foos = Foo.where(conditions).order(:id)

When params[:foo] is nil, this would call:

    foos = Foo.where(nil).order(:id)

In this scenario, we want Foo.where(conditions) to be the same as calling
Foo.all, otherwise we'd get a "NoMethodError order for WhereChain".

Related to #8332.

Relation.where with no args can be chained with not, like, and not_like

Conflicts:
	activerecord/CHANGELOG.md
	activerecord/lib/active_record/relation/query_methods.rb

When applying default_scope to a class with a where clause, using
update_column(s) could generate a query that would not properly update
the record due to the where clause from the default_scope being applied
to the update query.

    class User < ActiveRecord::Base
      default_scope where(active: true)
    end

    user = User.first
    user.active = false
    user.save!

    user.update_column(:active, true) # => false

In this situation we want to skip the default_scope clause and just
update the record based on the primary key. With this change:

    user.update_column(:active, true) # => true

Fixes #8436.

LOCALTIMESTAMP is not support by sqlite3, and travis was giving us these
errors: https://travis-ci.org/rails/rails/jobs/3535241/#L570

Before:

    Finished tests in 56.245787s, 0.2133 tests/s, 0.0000 assertions/s.
    12 tests, 0 assertions, 0 failures, 0 errors, 0 skips

After:
    Finished tests in 42.401416s, 0.2830 tests/s, 0.0000 assertions/s.
    12 tests, 0 assertions, 0 failures, 0 errors, 0 skips

CHANGELOG for 78d5d6f8

[ci skip]

It was pointed out by @giner that the CHANGELOG entry for https://github.com/rails/rails/commit/78d5d6f8688bb7c45ba9a3ef893682231130da3f
wasn't included. Here it is.

Fixes issues in test_mysql2

This commit inverts the precedence in ActionDispatch::Static so that
dynamic content will be served before static content. This is so that
precompiled assets do not inadvertently get included when running in
development mode - it should have no effect in production where static
files are usually handled by the web server.

Closes #6421

why is this a Time to start with?

This cleanup aims to fix a build failure:
https://travis-ci.org/rails/rails/jobs/3515951/#L482

Since travis always have both ENV vars set to "test", a test is failing
where it's expected to output the default env "development", but "test"
is the result due to RACK_ENV being set when we expect it to not be.

By cleaning this duplication we ensure that changing any of these env
variables will pick the right expected value.

To prevent future pull requests like #8435 add a comment about the
implementation of class_attribute using class_eval for performance.

[ci skip]

dump schema.rb without :version option

* joshsusser-master:
  style cleanup
  Add migration history to schema.rb dump
  Add metadata to schema_migrations

Conflicts:
	activerecord/CHANGELOG.md
	activerecord/lib/active_record/schema.rb

Revert turbolinks load order requirement

[ci skip]

Add ENV['RACK_ENV'] support to rake runner/console/server.

fix warning: possibly useless use of a constant in void context

Allow fragment cache to accept :if and :unless options.

Closes #5396

[Stephen Ausman + Fabrizio Regini]

Extract method refactoring.

Replace comments' non-breaking spaces with spaces [ci skip]