| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
| |
ActionDispatch::ParamsParser class was removed in favor of
ActionDispatch::Http::Parameters so it is better to move the error
constant to the new class.
|
| |
|
| |
|
| |
|
|
|
|
|
| |
assigns assert the state of a controller instance what should not be
done in an integration test.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently a misleading "missing required keys" error is thrown when a param
fails to match the constraints of a particular route. This commit ensures that
these params are recognised as unmatching rather than missing.
Note: this means that a different error message will be provided between
optimized and non-optimized path helpers, due to the fact that the former does
not check constraints when matching routes.
Fixes #26470.
|
|
|
|
|
|
|
|
|
|
| |
When passed an already-valid file name, prepending the path is likely to
create problems.
This is particularly relevant for #26384, which adds fixture_path
handling to test classes that previously didn't have it: any existing
caller must have been manually locating the file, and we don't want to
break them.
|
|\
| |
| | |
Fix memoization bug on ActionDispatch::TestRequest#request_method=
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
TestRequest have been overrriding request_method setter since 2009,
but the actual implementation in Request (not TestRequest) has been
changed since that. Now it's also using @request_method instance
variable to keep the state.
The override in TestRequest have not been calling `super`, which caused
a bug that after accessing #requst_method the value was memoized and
then we've never been able to change it anymore:
```
req = ActionDispatch::TestRequest.create
puts "was: #{req.request_method}" # memoized here
req.request_method = "POST"
puts "became: #{req.request_method}"
```
output:
```
was: GET
became: GET
```
Since the whole purpose of overriding the setter in TestRequest is to
upcase it, I'm changing it to `super(method.to_s.upcase)`
|
|\ \
| | |
| | | |
Improve assert_response helper
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When the check is failed, print the actual response body if it's not too large.
This could improve productivity when writing new tests.
Before:
```
ThemeEditorIntegrationTest#test_whatever
Expected response to be a <200: ok>, but was a <422: Unprocessable Entity>.
Expected: 200
Actual: 422
```
After:
```
ThemeEditorIntegrationTest#test_whatever
Expected response to be a <200: ok>, but was a <422: Unprocessable Entity>.
Expected: 200
Actual: 422
Response body: {"errors":["Invalid settings object for section '1'"]}
```
|
|/
|
|
|
|
| |
All indentation was normalized by rubocop auto-correct at 80e66cc4d90bf8c15d1a5f6e3152e90147f00772.
But comments was still kept absolute position. This commit aligns
comments with method definitions for consistency.
|
|\
| |
| |
| |
| | |
y-yagi/make_fixture_file_upload_in_integration_test
make `fixture_file_upload` work in integration tests
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Currently, `fixture_file_upload` does not work in integration test.
Because, `TestProcess` module has been include in `Session` class, but
`fixture_path` can not get from `Session` class.
Modify to include `TestProcess` in `IntegrationTest` class in order to get
correct value of `fixture_path`.
|
|/
|
|
|
|
| |
All indentation was normalized by rubocop auto-correct at 80e66cc4d90bf8c15d1a5f6e3152e90147f00772.
But heredocs was still kept absolute position. This commit aligns
heredocs indentation for consistency.
|
| |
|
| |
|
|
|
|
| |
key length
|
|
|
|
|
|
|
|
|
| |
Since keys are truncated, ruby 2.4 doesn't accept keys greater than their lenghts.
keys of same value but different lenght and greater than key size of cipher, produce the same results
as reproduced at https://gist.github.com/rhenium/b81355fe816dcfae459cc5eadfc4f6f9
Since our default cipher is 'aes-256-cbc', key length for which is 32 bytes, limit the length of key being passed to Encryptor to 32 bytes.
This continues to support backwards compat with any existing signed data, already encrupted and signed with 32+ byte keys.
Also fixes the passing of this value in multiple tests.
|
|
|
|
|
|
|
|
|
|
| |
This was almost every case where we are overriding `respond_to?` in a
way that mirrors a parallel implementation of `method_missing`. There is
one remaining case in Active Model that should probably do the same
thing, but had a sufficiently strange implementation that I want to
investigate it separately.
Fixes #26333.
|
|\
| |
| | |
Allow `send_file` to declare a charset
|
| | |
|
| |
| |
| |
| |
| |
| | |
Removed my patch in favor of @tenderlove's less invasive approach.
[Aaron Patterson & Jon Moss]
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The PR #20940 enabled the use of multiple roots with different constraints
at the top level but unfortunately didn't work when those roots were inside
a namespace and also broke the use of root inside a namespace after a top
level root was defined because the check for the existence of the named route
used the global :root name and not the namespaced name.
This is fixed by using the name_for_action method to expand the :root name to
the full namespaced name. We can pass nil for the second argument as we're not
dealing with resource definitions so don't need to handle the cases for edit
and new routes.
Fixes #26148.
|
|
|
|
|
|
| |
It should not throw a NameError, but should throw a KeyError.
Fixes #26278
|
|\
| |
| |
| |
| | |
samphippen/allow-early-setting-of-integration-session
Allow the `integration_sesion` to be set early on ActionDispatch::Integration::Runner.
|
| |
| |
| |
| | |
This allows us to not `||=` in `before_setup`.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
ActionDispatch::Integration::Runner.
In commit fa63448420d3385dbd043aca22dba973b45b8bb2, @tenderlove changed
the behaviour of the way `integration_session` is set up in this object.
It used to be the case that the first time it was accessed, it was
memoized with nil, however, this means that if it had already been set
it was not replaced. After that commit, it is now always set to `nil` in
the execution of `before_setup`.
In RSpec, users are able to invoke `host!` in `before(:all)` blocks,
which execute well before `before_setup` is ever invoked (which happens
in what is equivalent to a `before(:each)` block, for each test. `host!`
causes the integration session to be set up to correctly change the
host, but after fa63448420d3385dbd043aca22dba973b45b8bb2 the
`integration_session` gets overwritten, meaning that users lose their
`host!` configuration (see https://github.com/rspec/rspec-rails/issues/1662).
This commit changes the behaviour back to memoizing with `nil`, as
opposed to directly overwriting with `nil`. This causes the correct
behaviour to occur in RSpec, and unless I'm mistaken will also ensure
that users who want to modify their integration sessions early in rails
will also be able to do so.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When `config.force_ssl` is set to `true`, any POST/PUT/DELETE requests coming in to non-secure url are being redirected with a 301 status.
However, when that happens, the request is converted to a GET request and ends up hitting a different action on the controller.
Since we can not do non-GET redirects, we can instead redirect with a 307 status code instead to indicate to the caller that a fresh request should be tried preserving the original request method.
`rack-ssl` gem which was used to achieve this before we had this middleware directly baked into Rails also used to do the same, ref: https://github.com/josh/rack-ssl/blob/master/lib/rack/ssl.rb#L54
This would be specially important for any apps switching from older version of Rails or apps which expose an API through Rails.
|
|/
|
|
| |
Use ActionDispatch::Request instead of Request because ActionDispatch::Request no longer inherits from Rack::Request.
|
|
|
|
|
| |
Those methods are only using inside this module and by a private method
so they all should be private.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since e852daa6976cc6b6b28ad0c80a188c06e226df3c only the verb methods
where extracting the defaults options. It was merged a fix for the
`root` method in 31fbbb7faccba25b2e3b5e10b8fca1468579d629 but `match`
was still broken since `:defaults` where not extracted.
This was causing routes defined using `match` and having the `:defaults`
keys to not be recognized.
To fix this it was extracted a new private method with the actual
content of `match` and the `:defaults` extracting was moved to `match`.
|
|\
| |
| | |
fix Rails.application.routes.router.visualizer for router debugging
|
| |
| |
| |
| | |
fixes error due to Routes#partitioned_routes being removed
|
| |
| |
| |
| |
| |
| |
| |
| | |
Style/SpaceBeforeBlockBraces
Style/SpaceInsideBlockBraces
Style/SpaceInsideHashLiteralBraces
Fix all violations in the repository.
|
| |
| |
| |
| |
| |
| | |
It's tough for people without the knowledge of where the `get` and
friends integration test helpers are defined to find documentation
for them. Add a link to the main integration test documentation.
|
|/
|
|
|
|
|
| |
* Give the section a header to distinguish it from the general doc.
* Replace backticks with + signs to fit SDoc.
* Use double quoted strings.
* Clarify how `parsed_body` works — it doesn't depend on `as` anymore.
|
| |
|
|
|
| |
When initializing an `ActionDispatch::Http::Headers` object it takes a request object (Rails 5) whereas before it took a hash (Rails 4.x) but the documented example still shows a hash given to the constructor (due to commit 34fa6658dd1b779b21e586f01ee64c6f59ca1537) so this is just a documentation change to use the new `from_hash` method introduced in that earlier commit.
|
|
|
|
|
| |
At GitHub we need to handle parameter encodings that are not UTF-8. This
patch allows us to specify encodings per parameter per action.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In integration test when specify the "Accept" header with "xhr: true"
option, the Accept header is overridden with a default xhr Accept
header. The issue only affects HTTP header "Accept" but not CGI variable
"HTTP_ACCEPT".
For example:
get '/page', headers: { 'Accept' => 'application/json' }, xhr: true
# This is WRONG! And the response.content_type is also affected.
# It should be "application/json"
assert_equal "text/javascript, text/html, ...", request.accept
assert_equal 'text/html', response.content_type
The issue is in `ActionDispatch::Integration::RequestHelpers`. When
setting "xhr: true" the helper sets a default HTTP_ACCEPT if blank.
But the code doesn't consider supporting both HTTP header style and
CGI variable style.
For detail see this GitHub issue:
https://github.com/rails/rails/issues/25859
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
The current code base is not uniform. After some discussion,
we have chosen to go with double quotes by default.
|