We were using `prepend: true`, but it doesn't seem to be necessary.
We were also loading the draft versions twice in the index, so we can
remove the line loading them a second time.
We forgot to remove it in commit 8066b96fe. We're removing it now
because these methods didn't follow the Layout/BlockAlignment rubocop
rule, which we're about to add.
For the HashAlignment rule, we're using the default `key` style (keys
are aligned and values aren't) instead of the `table` style (both keys
and values are aligned) because, even if we used both in the
application, we used the `key` style a lot more. Furthermore, the
`table` style looks strange in places where there are both very long and
very short keys and sometimes we weren't even consistent with the
`table` style, aligning some keys without aligning other keys.
Ideally we could align hashes to "either key or table", so developers
can decide whether keeping the symmetry of the code is worth it in a
case-per-case basis, but Rubocop doesn't allow this option.
We were already applying these rules in most cases.
Note we aren't enabling the `MultilineArrayLineBreaks` rule because
we've got places with many elements whire it isn't clear whether
having one element per line would make the code more readable.
Since IRB has improved its support for multiline, the main argument
towars using a trailing dot no longer affects most people.
It still affects me, though, since I use Pry :), but I agree
leading dots are more readable, so I'm enabling the rule anyway.
We added the `user_id` rule in commit edaf420f5. To be honest, I'm not
sure what we meant, since I haven't found URLs containing the user id.
So we're treating it as if it was a typo and we wanted to do the same
thing we did with other parameters.
In the `i18n_translation` initializer, we're overwriting the `t` helper
so calling it uses custom translations if they're available.
However, ViewComponent doesn't use the `t` helper but implements its own
`t` method. So, when calling the `t` method in a component, we weren't
using our implementation of the `t` helper, and so we weren't loading
custom translations.
Using the `t` helper in components solves the issue.
There was a test where we were directly testing a method in a component,
and that method uses the `t` helper. This caused an error when running
the test:
ViewComponent::Base::ViewContextCalledBeforeRenderError:
`#helpers` can't be used during initialization, as it depends on the
view context that only exists once a ViewComponent is passed to the
Rails render pipeline.
Using `render_inline` in the test and testing the generated HTML, as
recommended in the ViewComponent documentation, solves the issue.
It was obvious which project the comment was about. Also, no point
having a TODO referencing what to do after upgrading to Rails 5.1, since
we did it years ago.
We were using two different sets of extensions but, since the markdown
code is always written by administrators, IMHO it makes sense to be
consistent and always render markdown code the same way.
* Add Tables option to Redcarpet in Legislation draft
* Allow table tags in Admin Legislation Sanitizer
* Add Test to render markdown tables in Legislation drafts
* Add Test for Admin Legislation Sanitizer
We include test for image, table and h1 to h6 tags and additional tests to strengthen the allowed and disallowed parameters
* Add Table from markdown test in System and Factories
* Add test to render tables for admin user
* Remove comment line about Redcarpet options
* Edit custom css for legislation draft table to make it responsive
The `inverted-selection` rules defined in the `%brand-background`
selector weren't being applied because we were using this selector in
`::before` and `::after` pseudoelements. Not sure about the reason, but
it looks like the saturation of `::after::selection` pseudoelements
resulted in invalid selectors and so the inverted selection rules were
ignored for every selection using `%brand-background`, like
`%budget-header`.
Using `@include brand-background` instead of `@extend %brand-background`
in pseudoelements solves the issue. The inverted selection might not
work in these pseudoelements, but we don't need it there since these
pseudoelements don't have content.
Zooming with the mousewheel is useful when you want to use it, but
annoying when you don't want to.
Here we're taking an intermediary approach: by default, the mousewheel
isn't active, but it will be active after focusing on the map, so it can
be used both to scroll and to zoom.
This behavior presents usability issues, though, since we aren't making
users aware of the way the mousewheel works, and even if they were
aware, it could be confusing anyway. However, I currently think it's
better than always enabling or always disabling the mousewheel (might
change my mind, though).
Note that the "focus" event is only used on the map, so if we click on a
marker or navigate to a marker with the keyboard without focusing on the
map first, the mousewheel isn't enabled. The same would happen if we
used the "click" event.
We might use the Leaflet.GestureHandling plugin in the future to deal
with this issue and the scroll on touch screens.
Just like we did for budgets, we're doing the same thing in all the
places where we render background images attached by either regular
users or administrators.
This way we correctly render background images with characters like
brackets or quotes.
In these commits 38ba5e159b and 8805037e2f we added the parameter
"vote_url" in the call to the partial "votes" in collaborative legislation
proposals.
It seems that this parameter is only used in the proposals module and not in
collaborative legislation proposals.
While it is true that in the partial "votes" in these commits this parameter
"vote_url" is referred to, in commit 276baedcf it seems to be removed.
For this reason, we remove the parameter.
By using the bindPopup function instead of the click event
popups work when using the keyboard.
Note that now we are loading all the map markers in the first
request in a single query to the database (needed when there
is a lot or markers to show). Because of that we removed the
AJAX endpoint.
This methods performs much better when we need to load a lot of
globalized models translations and returns the best fallback translation
for the current language.
Images with brackets weren't being rendered properly, so we're now
enclosing the URL in single quotes. In order to render images with
single quotes, we're also using the `j` method.
Since the tests were written by a different developer, we're adding them
in the next commit.
Previous to this commit the geozone link shown in the
legislation proposal page was pointing to the proposals
process feature instead to the legislation proposals.
