* The method name must be `stored_attributes`, not `stores_attributes`.
* `attribute_names` must return a non-empty value. Because
`stored_attributes` is not checked if `attribute_names` is empty.
Follow up to #28056
This reverts commit c6f9f8c28a, reversing
changes made to c309073c74.
Reason: This is fixing the behavior in the wrong place. Now the request
path after the request is nil and there is no way to assert that.
Also the test that was added in that PR also fails in 4.2 where the
reporter says it was passing. The reason the bahavior changed between
Rails 4.2 and Rails 5 is that the format in the path is now respected.
The correct way to fix the problem is not doign two requests in the same
controller test and use integrations tests. This change caused a
regression between Rails 5.0.1 and 5.0.2.
Action View overrides `url_for` in the view context to render paths by
default when using `url_for` and this means that direct route helpers
don't get the full url when called with the url suffix. To fix this
always call the original `url_for`.
Not all requirements can be expressed in terms of polymorphic url
options so add a `route_for` method that allows calling another
direct route (or regular named route) which a set of arguments, e.g:
resources :buckets
direct :recordable do |recording|
route_for(:bucket, recording.bucket)
end
direct :threadable do |threadable|
route_for(:recordable, threadable.parent)
end
This maintains the context of the original caller, e.g.
threadable_path(threadable) # => /buckets/1
threadable_url(threadable) # => http://example.com/buckets/1
Capybara was updated in teamcapybara/capybara#1841 to use Minitest style
assertions so that system test output shows x number of assertions, x
numbe of failures, etc.
Before:
```
6 runs, 0 assertions, 0 failures, 0 errors, 0 skips
```
After:
```
6 runs, 7 assertions, 1 failures, 0 errors, 0 skips
```
This change bumps Capybara from 2.7.0 to 2.13.0 and includes the
required minitest assertion file in the test case. 🎉
Looking on code for this method it's clear that it's just returns `response.status` instead of full `response` object.
It's better to correct docs as probably lots of specs are relying on this behavior.
In 9b654d4 some params munging was added to ensure that they were
set whenever `recognize_path` would call either a proc or callable
constraint. Since we no longer mutate the environment hash within
the method it's now unnecessary and actually causes params to leak
between route matches before checking constraints.
Fixes#28398.
Capybara drivers can handle some options such like `url`.
### before
```
# test/test_helper.rb
Capybara.register_driver :remote_chrome do |app|
Capybara::Selenium::Driver.new(app, browser: :chrome, url: "http://example.com/wd/hub")
end
# test/application_system_test_case.rb
class ApplicationSystemTestCase < ActionDispatch::SystemTestCase
driven_by :remote_chrome
end
```
### after
```
# test/application_system_test_case.rb
class ApplicationSystemTestCase < ActionDispatch::SystemTestCase
driven_by :selenium, using: :chrome, screen_size: [1400, 1400], options: {url: "http://chrome:4444/wd/hub"}
end
```
3 output types are supported:
- simple: only display the screenshot path
- artifact: display the screenshot in the terminal, using the artifact protocol (supported by some CI)
- inline (default): display the screenshot in the terminal, inline (supported by some terminals)
You can force the output type by setting the `RAILS_SYSTEM_TESTING_SCREENSHOT` environment variable
Previously the system test subclasses would call `driven_by` when the
app booted and not again when the test was initialized which resulted in
the driver from whichever class was called last to be used in tests.
In rails/rails#28144 the `driven_by` method was changed to run `use` on
setup and `reset` on teardown. While this was a viable fix this really
pointed to the problem that system test `driven_by` was a global
setting, rather than a per-class setting.
To alieviate this problem calling the driver should be done on an
instance level, rather than on the global level. I added an `initialize`
method to `SystemTestCase` which will call `use` on the superclass
driver. Running the server has been moved to `start_application` so that
it only needs to be called once on boot and no options from `driven_by`
were being passed to it.
This required a largish rewrite of the tests. Each test needs to utilize
the subclass so that it can properly test the drivers.
`ActionDispatch::SystemTestCase` shouldn't be called directly anymore.
Since using a browser is only for selenium it doesn't really make sense
to have a separate class for handling it there. This brings a lot of the
if/else out of the main SystemTestCase class and into the Driver class
so we can abstract away all that extra work.
Since `ActionController:Parameters` does not inherit `Hash`, need to
explicitly convert it to `Hash`.
Also, `Parameters#to_h` returns `Hash` whose key is `String`. Therefore,
if merge as it is, the value will not be overwritten as expected.
```diff
diff --git a/actionpack/lib/action_controller/metal/redirecting.rb b/actionpack/lib/action_controller/metal/redirecting.rb
index a3159c29dd..1836a07d4e 100644
--- a/actionpack/lib/action_controller/metal/redirecting.rb
+++ b/actionpack/lib/action_controller/metal/redirecting.rb
@@ -50,8 +50,8 @@ module Redirecting
# redirect_to post_url(@post), status: 301, flash: { updated_post_id: @post.id }
# redirect_to({ action: 'atom' }, alert: "Something serious happened")
#
- # Statements after redirect_to in our controller get executed, so redirect_to doesn't stop the execution of the function.
- <U+2028># To terminate the execution of the function immediately after the redirect_to, use return.
+ # Statements after +redirect_to+ in our controller get executed, so +redirect_to+ doesn't stop the execution of the function.
+ # To terminate the execution of the function immediately after the +redirect_to+, use return.
# redirect_to post_url(@post) and return
def redirect_to(options = {}, response_status = {})
raise ActionControllerError.new("Cannot redirect to nil!") unless options
```
https://travis-ci.org/rails/rails/jobs/207908041#L549
This reverts commit 0d73f9116c, reversing
changes made to 21ff8a4932.
We don't want to silence Puma because it's start up messages can be
useful (to see threads, port etc). #28109 was instead fixed by not
running system tests by defaul with `bin/rails test` in #28286.
Should be `ActionController::ParameterMissing` and not
`ActionController::MissingParameter`.
Corresponding change was done in guides in
https://github.com/rails/rails/pull/9816.
`Capybara::RackTest::Driver` does not support taking screenshots. If call
`#save_screenshot` on `Capybara::RackTest::Driver` will raise the error.
```ruby
Error:
UsersTest#test_visiting_the_index:
Capybara::NotSupportedByDriverError: Capybara::Driver::Base#save_screenshot
```
To prevent errors, if driver does not support screenshot, do not call it.
If application has ajax, browser may begin request after rollback.
`teardown_fixtures` will be called after `super` on `after_teardown`
so we must call `Capybara.reset_sessions!` before `super`
b61a56541a/activerecord/lib/active_record/fixtures.rb (L857)
Now reset the driver before take failed screenshot since #28144.
However, I think that failed screenshot should be take with the driver
actually used in the test.
So, fixed to take screenshot before reset driver.
In ca324a0 the flash middleware was effectively removed by its
constructor returning the app it was passed and the `commit_flash`
call was moved to the `ActionController::Metal#dispatch` method.
This broke any redirect routes that modified the flash because the
redirect happens before `dispatch` gets called.
To fix it, this commit adds a `commit_flash` call in the `serve`
method of `ActionDispatch::Routing::Redirect`.
Fixes#27992.
A gzip file has a checksum and length for the decompressed data in its
footer which isn't checked by just calling Zlib::GzipReader#read.
Calling Zlib::GzipReader#close must be called after reading to the end
of the file causes this check to be done, which is done by
Zlib::GzipReader.wrap after its block is called.
Previously, `driven_by` would change the Capybara configuration when
the test case is loaded, and having multiple test classes with different
`driven_by` configs would fail as the last loaded would be effective.
Since a `direct` url helper block is evaluated using `instance_exec`
then methods that are available in the instance context can be
accessed, e.g. the params object in a controller action or view.
This wasn't clear from the example so expand on that point and add
a test case for this situation.
Using `undef_method` means that when a route is removed any other
implementations of that method in the ancestor chain are inaccessible
so instead use `remove_method` which restores access to the ancestor.
Allow the use of `direct` to specify custom mappings for polymorphic_url, e.g:
resource :basket
direct(class: "Basket") { [:basket] }
This will then generate the following:
>> link_to "Basket", @basket
=> <a href="/basket">Basket</a>
More importantly it will generate the correct url when used with `form_for`.
Fixes#1769.
Allow the definition of custom url helpers that will be available
automatically wherever standard url helpers are available. The
current solution is to create helper methods in ApplicationHelper
or some other helper module and this isn't a great solution since
the url helper module can be called directly or included in another
class which doesn't include the normal helper modules.
Reference #22512.
The singleton url_for on Rails.application.routes.url_helpers isn't the
same as the url_for you get when you include the module in your class as
the latter has support for polymorphic style routes, etc. whereas the
former accepts only a hash and is the underlying implementation defined
on ActionDispatch::Routing::RouteSet.
This commit changes the singleton method to call through a proxy instance
so that it gets the full range of features specified in the documentation
for url_for.
This renames the system test helper file to be application system test
case to match what the rest of Rails does. In the future we should
consider changing the test_helper to match.
* Override integration test default host
Integration tests automatically set the default host to
'http://example.com'. This works fine for integration tests because they
are not real browser sessions, but doesn't work fine for system tests
because they are real browser sessions.
We can override this by setting the `host!` in `before_setup. The
`Capybara.always_include_port` will allow the test to look at
`127.0.0.1:port capybara picks` and properly redirect the test.
Any application can override this by setting the `host!` in
their system test helper. Generally though, applications are going to be
using localhost.
In this commit I also moved the setup and teardown into their own module
for tidiness.
* Move teardown settings into system test case
These configuration options can be put into the system test case file
instead of the generated system tests helper file. This is an
implementation detail and therefore shouldn't be generated with the
template.
* Move system tests back into Action Pack
* Rename `ActionSystemTest` to `ActionDispatch::SystemTestCase`
* Remove private base module and only make file for public
`SystemTestCase` class, name private module `SystemTesting`
* Rename `ActionSystemTestCase` to `ApplicationSystemTestCase`
* Update corresponding documentation and guides
* Delete old `ActionSystemTest` files
Renames `Rails::SystemTestCase` to `ActionSystemTest` and moves it to a
gem under the Rails name.
We need to name the class `ActionSystemTestCase` because the gem expects
a module but tests themselves expect a class.
Adds MIT-LICENSE, CHANGELOG, and README for the future.
Puma is the default webserver of Rails. Because of this it doesn't make
sense to run tests in Webkit if the default server is Puma.
Here I've refactored the webserver to be it's own standalone module so
it can be shared between Rails' selenium default driver and Capybara's
defaut drivers.
Jon Moss