This reverts commit ade756fe.

Valid hex strings aren't valid float column values, to match the integer restriction.  [#4622 state:resolved]

[#4612]
Signed-off-by: NJeremy Kemper <jeremy@bitsweat.net>

[#2919 state:committed]
Signed-off-by: NJeremy Kemper <jeremy@bitsweat.net>

[#3642 state:committed]
Signed-off-by: NMarius Nuennerich <marius@nuenneri.ch>
Signed-off-by: NJeremy Kemper <jeremy@bitsweat.net>

a cloned object no longer mimics changed flags from creator , plus a test case [#4614 state:resolved]
Signed-off-by: NJosé Valim <jose.valim@gmail.com>

This reverts commits af0d1a88 and 64d109e3.

Signed-off-by: NXavier Noria <fxn@hashref.com>

* Default Encoding.default_internal to UTF-8
* Eliminated the use of file-wide magic comments to coerce code evaluated inside the file
* Read templates as BINARY, use default_external or template-wide magic comments
  inside the Template to set the initial encoding
  * This means that template handlers in Ruby 1.9 will receive Strings encoded
    in default_internal (UTF-8 by default)
* Create a better Exception for encoding issues, and use it when the template
  source has bytes that are not compatible with the specified encoding
* Allow template handlers to opt-into handling BINARY. If they do so, they
  need to do some of their own manual encoding work
* Added a "Configuration Gotchas" section to the intro Rails Guide instructing
  users to use UTF-8 for everything
* Use config.encoding= in Ruby 1.8, and raise if a value that is an invalid
  $KCODE value is used

Also:
* Fixed a few tests that were assert() rather than assert_equal() and
  were caught by Minitest requiring a String for the message
* Fixed a test where an assert_select was misformed, also caught by
  Minitest being more restrictive
* Fixed a test where a Rack response was returning a String rather
  than an Enumerable

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

prevent to run fixture accessor (e.g. test_foos for TestFoo model) as a test case [#2992 state:resolved]
Signed-off-by: NWijnand Wiersma <wijnand@videre.net>
Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NXavier Noria <fxn@hashref.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

[#4610 state:committed]
Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Make use of assert_equal to test equallity between object assert expects and object and a message of error

[#4611 state:committed]
Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Postgresql doesn't allow to change a string type column to a binary type. Skip this test for postgresql for now. [#4616 state:resolved]
Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

using :time_select when the attribute type is :time in the scaffold generator. [#2377 state:resolved]
Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Ensure that eager_load actually takes place just after the middleware stack is built by using another pattern.

Also create a engine_blank_point initializer to ensure any :before or :after hooks defined inside engines won't move the configuration initializers to other places.

Signed-off-by: NJosé Valim <jose.valim@gmail.com>

Revert "Make ActiveModel::Errors#add_on_blank and #add_on_empty accept an options hash and make various Validators pass their (filtered) options."

Having a huge array to whitelist options is not the proper way to handle this case. This means that the ActiveModel::Errors object should know about the options given in *all* validators and break the extensibility added by the validators itself. If the intent is to whitelist options before sending them to I18n, each validator should clean its respective options instead of throwing the responsibility to the Errors object.

This reverts commit bc1c8d58.