American English(according to Weber)

That integer is rather irrelevant, the only thing that matters is that
it is consistent and with no apparent bias. Zlib.crc32 is 8-10 times
faster than bytes.sum, so use that.

Using the sum of bytes instead the hash of the path when replacing the wildcard of the assets path because in ruby 1.9 is not consistent

  config.action_controller.default_asset_host_protocol

It's best to leave this unset.

When unset the :request protocol is used whenever it can be and
:relative is used in the other situations.

When set to :request then assets hosts will be disabled when there
is no request in scope and will use the request protocol whenever a
request is in scope.

If set to :relative, then a relative protocol is always used except
for stylesheet link tags which must use the :request protocol to
avoid double downloads in IE6&7.

Conflicts:

	actionpack/lib/sprockets/helpers/rails_helper.rb
	actionpack/test/template/sprockets_helper_test.rb

Stylesheet link tags should use the request protocol to avoid duplicate download of stylesheets in IE7 and IE8.

Conflicts:

	actionpack/lib/action_view/helpers/asset_tag_helpers/stylesheet_tag_helpers.rb
	actionpack/lib/sprockets/helpers/rails_helper.rb

Add asset_url helper and refactor the asset paths so that asset hosts can be used during asset precompilation.

Conflicts:

	actionpack/lib/action_view/asset_paths.rb
	actionpack/lib/sprockets/helpers/rails_helper.rb
	actionpack/test/template/sprockets_helper_test.rb

This reverts commit 057412ce.

Applied changes to stylesheet_link_tag from javascript_include_tag which corrects issues with ordering and duplicates.

This corrects two issues with javascript_include_tag, the order at which they are expanded, and removing duplicates.

When individual js assets are specified, they will override the order of the same asset specified in an expansion.

[#5938 state:resolved]

Using the same hash instance makes using the same expansions for
both javascripts and stylesheets.

Signed-off-by: NSantiago Pastorino <santiago@wyeworks.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>

adds #favicon_link_tag back, rdoc explains why it is useful, and how to get a link for Mobile Safari with it

This reverts commit 6891f46d, ec8610cf, and  d18ff1b7.

Refactored url_for in AV to have its own instances of the helpers instead of proxying back to the controller. This potentially allows for more standalone usage of AV. It also kicked up a lot of dust in the tests, which were mocking out controllers to get this behavior. By moving it to the view, it made a lot of the tests more standalone (a win) 

Set config.action_controller.asset_path_template to a proc that takes one
argument - the direct, unchanged asset path - and returns a path to that asset
using the scheme that your assets require.

This is useful if you have a setup which scales by introducing new application
servers where the mtime of the asset files may not be the same as those of the
asset files on your previous servers, but it does require your web servers to
have knowledge of the asset template paths that you rewrite to so it's not
suitable for out-of-the-box use. An example of configuring asset path
generation and rewriting these paths using Apache is included in
actionpack/lib/action_view/helpers/asset_tag_helper.rb.
Signed-off-by: Nwycats <wycats@gmail.com>

There's a Ruby issue with File.basename different versions returns different things, so we shouldn't test that