| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| |\
| |
| | |
replace regexp global in #url_for
|
| | | |
|
| |\ \
| | |
| | | |
Improve img alt attribute for screen readers
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Currently, the img_alt method in ActionView keeps underscores
in the alt attribute. Because underscores are pronounced in
Apple's VoiceOver Utility, this has serious implications for
accessibility. This patch makes underscored or hyphenated file
names (both common in projects) read more naturally in screen
readers by replacing them with spaces. See method documentation
for details.
Added documentation to image_alt method
|
| | | | |
|
| |/ / |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The current implementation only works correctly if you supply the `:controller`
with directory notation (eg. `:controller => 'admin/posts'`).
The ruby constant notation (eg. `:controller => 'Admin::Posts`) leads to unexpected problems with `url_for`.
This patch prints a warning for every non supported `:controller` option. I also added documentation how
to work with namespaced controllers. The warning links to that documentation in the rails guide.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
previously when a partial was placed inside a directory
(eg. '/dir/_partial'), `assert_template` did not replace
the '_' prefix when looking through rendered tempaltes,
which resulted in an error.
I modified it to replace both, the leading '_' and the last '_'
after a '/'.
|
| | |
| |
| |
| |
| |
| |
| | |
When `assert_template` is used with the :locals option, and the
partial was not rendered, a method_missing error was raised.
This changes first checks, if the partial actually was rendered
and raises a descriptive error.
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
It would embed array as string instead of joining it like `content_tag` does:
content_tag(:td, class: ["foo", "bar"]){}
#=> '<td class="foo bar"></td>'
Before:
content_tag_for(:td, item, class: ["foo", "bar"]){}
#=> '<td class="item ["foo", "bar"]" id="item_1"></td>'
After:
content_tag_for(:td, item, class: ["foo", "bar"]){}
#=> '<td class="item foo bar" id="item_1"></td>'
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| | |
I don't believe this test is exercising any explicit params_parser
behavior that the other two Proc tests aren't already doing. Given
that we now know it's a bad idea to load user input via YAML.load,
somebody reading this test might get a dangerous idea about building
out a YAML params parser.
|
| | |
| |
| |
| |
| |
| | |
ActionDispatch::Head was removed in favor of Rack::Head. But Rack::Head
does not convert GET requests to HEAD requests so we need to do
checking for HEAD requests ourselves.
|
| |\ \
| | |
| | | |
Make HEAD work / convert to GET once more
|
| | | |
| | |
| | |
| | | |
protection
|
| | | | |
|
| | | | |
|
| | | |
| | |
| | |
| | | |
Bring back the same API we have with Request::Session.
|
| | | |
| | |
| | |
| | |
| | |
| | | |
All ActionPack and Railties tests are passing. Closes #8891.
[Carlos Antonio da Silva + Santiago Pastorino]
|
| | | |
| | |
| | |
| | | |
this would give us some more clues in case a test silently dies inside Thread
|
| |/ / |
|
| | | |
|
| | |
| |
| |
| | |
[ci skip]
|
| |\ \
| | |
| | |
| | | |
Now we're almost ready to remove this: https://github.com/rails/rails/blob/5294ad8/activesupport/lib/active_support/test_case.rb#L29
|
| | | | |
|
| | | | |
|
| | | | |
|
| |/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The [following pull request](https://github.com/rails/rails/pull/8916) fixed
the block being passed to the appropriate helper method. However, the content
being passed into the block is generating repeated markup on the page due to
some weird ERb evaluation.
This commit tries to capture the block's generated output so the page isn't
flooded with markup.
[Rafael França + José Mota]
Closes #8936
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Ruby 1.9 freezes Hash string keys by default so where a route is
defined like this:
get 'search' => 'search'
then the Mapper will derive the action from the key. This blows up
later when the action is added to the parameters hash and the
encoding is forced.
Closes #3429
|
| |\ \
| | |
| | | |
AR integration tests for form helpers
|
| | | | |
|
| |\ \ \
| | | |
| | | | |
In Browser Path Matching with Javascript
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
When debugging routes ,it can sometimes be difficult to understand exactly how the paths are matched. This PR adds a JS based path matching widget to the `/rails/info/routes` output. You can enter in a path, and it will tell you which of the routes that path matches, while preserving order (top match wins).
The matching widget in action:
Prior to this PR the only way to check matching paths is via mental math, or typing in a path in the url bar and seeing where it goes. This feature will be an invaluable debugging tool by dramatically decreasing the time needed to check a path match.
ATP actionpack
|
| | |/ /
|/| |
| | |
| | | |
It was being set to nil instead due to the wrong assignment.
|
| |\ \ \
| |/ /
|/| | |
Enhanced tests for AbstractController::Translation module
|
| | | |
| | |
| | |
| | | |
translation tests.
|
| | | | |
|
| | | | |
|
| |/ / |
|
| |\ \
| | |
| | | |
Remove i18n symbol dependency
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
date.order is the only key in rails i18n that is required to be a
symbol. This patch allows for symbols or strings which means:
* No requirement for symbol type in .yml files. A future
YAML.safe_load wouldn't need to load symbols
* Rails could actually use json rather than yml as the backend
|
| | | |
| | |
| | |
| | |
| | |
| | | |
https://github.com/rails/rails/commit/cc1c3c5be061e7572018f734e5239750ab449e3f
Now instead of raise, we log by default in development and test
|
| | | | |
|
| | | | |
|
| | | | |
|
| |/ / |
|
| | |
| |
| |
| | |
Also add some generic tests to ensure they're properly deprecated.
|
