| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The AV::Base constructor was too complicated, and this commit tightens
up the parameters it will take. At runtime, AV::Base is most commonly
constructed here:
https://github.com/rails/rails/blob/94d54fa4ab641a0ddeb173409cb41cc5becc02a9/actionview/lib/action_view/rendering.rb#L72-L74
This provides an AV::Renderer instance, a hash of assignments, and a
controller instance. Since this is the common case for construction, we
should remove logic from the constructor that handles other cases. This
commit introduces special constructors for those other cases.
Interestingly, most code paths that construct AV::Base "strangely" are
tests.
|
| |
| |
| |
| |
| | |
That method doesn't exist on LookupContext, so the delegate doesn't make
sense.
|
| |
| |
| |
| | |
We can remove the ivar by caching the digest on the stack
|
| |
| |
| |
| |
| |
| | |
This gets the PartialRenderer to be a bit closer to the
TemplateRenderer. TemplateRenderer already keeps its template in a
local variable.
|
| |
| |
| |
| | |
This method is private, and we always pass something in.
|
| |
| |
| |
| | |
Similar to 1853b0d0abf87dfdd4c3a277c3badb17ca19652e
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This reduces the surface area of our API and removes a Liskov issue.
Both TemplateRenderer and PartialRenderer inherit from AbstractRenderer,
but since PartialRenderer implements it's own `find_template` that is
private, and has the wrong method signature, an instance of
PartialRenderer cannot be substituted for an instance of
AbstractRenderer renderer. Removing the superclass implementation
solves both issues.
|
| |
| |
| |
| |
| |
| |
| | |
This patch changes `with_fallbacks` to be a factory method that returns
a new instance of a lookup context which contains the fallback view
paths in addition to the controller specific view paths. Since the
lookup context is more "read only", we may be able to cache them
|
| |
| |
| |
| |
| |
| | |
The `with_fallbacks` method will temporarily mutate the lookup context
instance, but nobody can call the setter, and we don't have to do a push
/ pop dance.
|
| |
| |
| |
| |
| |
| | |
We can't use the FixtureResolver as a hash key because it doesn't
implement `hash` correctly. This commit renames the method to "data"
(which is just as unfortunately named :( )
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The test passes an instance of `ActionController::Parameters` that acts
like a Hash but does not respond to some Hash methods like
`symbolize_keys`.
Moreover, if someone were to call `to_h` on the value it would fail since
the parameter is not permitted. So this is a great way to ensure that the
partial rendering pipeline does not mess with `locals`.
|
| | |
|
| |
| |
| |
| | |
Don't upsize images smaller than the specified dimensions.
|
| |
| |
| |
| |
| | |
If we pass the view instance around it's easier to understand the flow
control.
|
|\ \
| | |
| | | |
Template Handler Refactoring
|
| | |
| | |
| | |
| | |
| | |
| | | |
Rather than doing is_a? checks, ask the view object for its compiled
method container. This gives us the power to replace the method
container depending on the instance of the view.
|
| | |
| | |
| | |
| | |
| | | |
We always want to include this module. It'll be used in production
(maybe)
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This patch removes the instance writer of view_context_class.
Subclasses may override it, but it doesn't need to be written. This
also eliminates the need to cache the return value of the class level
`view_context_class` method.
|
| | | |
|
| | |
| | |
| | |
| | |
| | | |
Since everything goes through a `run` method, we can pull the buffer
assignment up.
|
| | |
| | |
| | |
| | |
| | | |
Methods created by views should always be evaluated against an AV::Base
instance. This way we can extract and refactor things in to classes.
|
| | |
| | |
| | |
| | |
| | |
| | | |
This pulls the "output buffer existence" conditional up. Instead of
evaling the same conditional over and over, we can pull it in to "only
compiled once" Ruby code.
|
| |/
|/|
| |
| |
| |
| | |
https://github.com/rails/rails/blob/v6.0.0.beta1/actionview/CHANGELOG.md
[ci skip]
|
| | |
|
| | |
|
|/ |
|
| |
|
|
|
|
| |
Fixes #33083
|
|
|
|
|
|
|
|
| |
yhirano55/rails_info_properties_json""
I reverted the wrong commit. Damn it.
This reverts commit f66a977fc7ae30d2a07124ad91924c4ee638a703.
|
|
|
|
|
|
|
|
|
|
|
| |
We had a discussion on the Core team and we don't want to expose this information
as a JSON endpoint and not by default.
It doesn't make sense to expose this JSON locally and this controller is only
accessible in dev, so the proposed access from a production app seems off.
This reverts commit 8eaffe7e89719ac62ff29c2e4208cfbeb1cd1c38, reversing
changes made to b6e4305c3bca4c673996d0af9db0f4cfbf50215e.
|
|\
| |
| |
| |
| | |
gsamokovarov/views-without-defined-protect-against-forgery
Don't expect defined protect_against_forgery? in {token,csrf_meta}_tag
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The `#csrf_meta_tags` and `#token_tag` Action View helper methods are
expecting the class in which are included to explicitly define the
method `#protect_against_forgery?` or else they will fail with
`NoMethodError`.
This is a problem if you want to use Action View outside of Rails
applications. For example, in #34788 I used the `#button_to` helper
inside of the error pages templates that have a custom
`ActionView::Base` subclass, which did not defined
`#protect_against_forgery?` and trying to call the button failed.
I had to dig inside of Action View to find-out what's was going on. I
think we should either set a default method implementation in the
helpers or check for the method definition, but don't explicitly require
the presence of `#protect_against_forgery?` in every `ActionViews::Base`
subclass as the errors are hard to figure out.
|
|/ |
|
|
|
|
|
|
|
|
| |
Because method arguments are different in the methods provided by form
helpers and form builders, I think these are necessary to prevent
confusion.
Fixes #34787
|
|\
| |
| | |
Enable `Style/RedundantBegin` cop to avoid newly adding redundant begin block
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Currently we sometimes find a redundant begin block in code review
(e.g. https://github.com/rails/rails/pull/33604#discussion_r209784205).
I'd like to enable `Style/RedundantBegin` cop to avoid that, since
rescue/else/ensure are allowed inside do/end blocks in Ruby 2.5
(https://bugs.ruby-lang.org/issues/12906), so we'd probably meets with
that situation than before.
|
|/
|
|
|
| |
* Fix integer regex deprecation warnings for Ruby 2.6.0
* Define =~ in FakeZone to avoid warnings from Ruby 2.6.0
|
|
|
|
|
|
| |
since Ruby 2.5
https://bugs.ruby-lang.org/issues/14133
|
|
|
|
|
|
|
|
|
|
| |
Generally followed the pattern for https://github.com/rails/rails/pull/32034
* Removes needless CI configs for 2.4
* Targets 2.5 in rubocop
* Updates existing CHANGELOG entries for fewer merge conflicts
* Removes Hash#slice extension as that's inlined on Ruby 2.5.
* Removes the need for send on define_method in MethodCallAssertions.
|
|\
| |
| | |
UJS: Do not disable previously disabled elements
|
| | |
|
| |
| |
| | |
[ci skip]
|
| |
| |
| |
| | |
[ci skip]
|
| | |
|
|/ |
|
|
|
|
|
| |
Firefox fires click events on left-, right-
and scroll-wheel (any non-primary mouse key) clicks while other browsers don't.
|
|
|
|
| |
side of long lines; Fixes #34487
|
|
|
|
| |
And remove trailing spaces.
|
|
|
|
|
|
|
| |
Reword first sentence of dep management and CVE section of
security guide. Also, reword and move gemspec notes above deps.
[ci skip]
|