Note that Rails 7.1 changes `find_or_create_by!` so it calls
`create_or_find_by!` when no record is found, meaning we'll rarely get
`RecordNotUnique` exceptions when using this method during a race
condition.
Adding this index means we need to remove the uniqueness validation.
According to the `create_or_find_by` documentation [1]:
> Columns with unique database constraints should not have uniqueness
> validations defined, otherwise create will fail due to validation
> errors and find_by will never be called.
We're adding a test that checks what happens when using
`create_or_find_by!`.
Note that we're creating voters combining `create_with` with
`find_or_create_by!`. Using `find_or_create_by!(...)` with all
attributes (including non-key ones like `origin`) fails when a voter
already exists with different values, e.g. an existing `origin: "web"`
and an incoming `"booth"`. In this situation the existing record is not
matched and the unique index raises an exception.
`create_with(...).find_or_create_by!(user: ..., poll: ...)` searches by
the unique key only and applies the extra attributes only on creation.
Existing voters are returned unchanged, which is the intended behavior.
For the `take_votes_from` method, we're handling a (highly unlikely, but
theoretically possible) scenario where a user votes at the same time as
taking voters from another user. For that, we're doing something similar
to what `create_or_find_by!` does: we're updating the `user_id` column
inside a new transaction (using a new transactions avoids a
`PG::InFailedSqlTransaction` exception when there are duplicate
records), and deleting the existing voter when we get a
`RecordNotUnique` exception.
On `Poll::WebVote` we're simply raising an exception when there's
already a user who's voted via booth, because the `Poll::WebVote#update`
method should never be called in this case.
We still need to use `with_lock` in `Poll::WebVote`, but not due to
duplicate voters (`find_or_create_by!` method will now handle the unique
record scenario, even in the case of simultaneous transactions), but
because we use a uniqueness validation in `Poll::Answer`; this
validation would cause an error in simultaneous transactions.
[1] https://api.rubyonrails.org/v7.1/classes/ActiveRecord/Relation.html#method-i-create_or_find_by
For the longest time, we've disabled the buttons to vote via web when
people had already voted in a booth. However, we were still allowing
HTTP requests to the actions to vote via web.
So we're adding a condition to prevent it.
The reason why we're changing the controller instead of the abilities
model (which is what we usually do) is that there might be side-effects
to the change. For instance, in the `Polls::PollComponent` class,
there's an `elsif cannot?(:answer, poll)` condition which would have a
different behavior if we changed the abilities model.
We were using similar code and three places. And all of them used
`15rem`, which looked a lot like a magic number. So we're making it the
default value of a mixin, which means replacing it with a less arbitrary
value will be easier.
Note that, for the column gap, we're now using the standard
grid-column-gutter we use in most places. We were using `$line-height`
in a couple of places only because writing it is less verbose.
Since we're now, for the first time, using a very long mixin definition
that we need to split in several lines, we're adding the `param`
exception to the indentation rule. As far as I know, there's no way to
define a rule in Stylelint that requires parameters in multiple lines to
be aligned with the first parameter, which is what we define in Rubocop.
Move resource cards layout inside #available-resources-section and switch
from equalizer alignment to a responsive grid layout.
Note that using `grid-auto-rows: 1fr` requires us to change the CSS of
the card itself so the "see resource" link remains at the bottom of the
card.
It's August 2025 and support for grid layout has been available in about
99% of the browsers for some time now. All major browsers added support
for grid layouts in 2017, which means our rule to support browsers that
are 7 years old allows us to start using `display: grid`.
Using a grid layout allows displaying a dynamic number of rows while
keepin all of them the same height, the same foundation's equalizer
does, by setting `grid-auto-rows: 1fr`.
And the `grid-template-columns` property lets us use dynamic columns for
all screen sizes, always filling the available space. No need to use
breakpoints.
The problem with "recommended" is that it had identical singular and
plural, which means it wasn't clear whether we were using it to refer to
an item of a collection or to refer to the collection itself.
We're also removing the padding in the (now called) followables element,
since it caused a gap on the right side of the border of the
`.menu.simple` element.
Note we aren't using flex-with-gap because currently (until we decide to
use the `gap` property) this mixin sets a negative margin that would
move the border of this element to the left.
This way we remove duplication in the HTML.
We're also adding a test checking what happens when users can vote in
order to test the `render?` method we've added.
Currently dependabot is failing to upgrade some gems that are part of
Rails. For example, when there's a security issue in ActiveRecord or
ActiveStorage, we get messages like:
```
Dependabot cannot update activestorage to a non-vulnerable version.
The latest possible version that can be installed is 7.1.5.1 because of
the following conflicting dependencies:
rails (7.1.5.1) requires activestorage (= 7.1.5.1) via actionmailbox (7.1.5.1)
rails (7.1.5.1) requires activestorage (= 7.1.5.1) via actiontext (7.1.5.1)
rails (7.1.5.1) requires activestorage (= 7.1.5.1)
The earliest fixed version is 7.1.5.2.
```
So we're relaxing the dependency in order to make it easier for
dependabot to upgrade gems that are part of Rails.
Note that, with this configuration, Dependabot wouldn't be able to
upgrade to Rails 7.1.6 if this releases fixed a security issues in a gem
that is part of Rails. We might still need to upgrade Rails manually in
this case.
This way polls look more similar to the way they did when the answers
were buttons instead of checkboxes or radio buttons.
Note the styling is tricky because we need to add a `float` property to
the legend so it's actually inside the fieldset. This forces us to add a
`::before` pseudo-element in order to add margin between the legend and
the first label. Another option would be:
```
legend {
&:has(+ label) {
margin-bottom: calc($line-height / 2);
}
+ label {
clear: $global-left;
}
}
```
But the `:has` pseudo-class isn't universally supported yet, and we'd
still have to add `margin-top` to the first label when it comes after a
`.help-text` element.
Due to the presence of the border, we're increasing the margin between
elements a little bit.
Note that adding a pseudoelement to the label is a consequence of adding
the `float` property to the legend, so we're changing the order of the
code so the styles for `legend` appear before the styles in `label`.
This could be the case when JavaScript is disabled.
Note that, in `Poll/WebVote` we're calling `given_answers` inside a
transaction. Putting this code before the transaction resulted in a test
failing sometimes, probably because of a bug that might be possible to
reproduce by doing simultaneous requests.
With the old interface, there wasn't a clear way to send a blank ballot.
But now that we've got a form, there's an easy way: clicking on "Vote"
while leaving the form blank.
Some of the code was in its own component, while some of the code
remained in the polls/show view.
Note that we're re-structuring the code a little bit, so it's clear that
the "already voted" messages are only shown when users can vote. Also
note that now the `can?` condition involves the existence of a
`current_user` and that the poll is not expired, so we can simplify the
`voted_in_web` condition.
This way it'll be easier to refactor it.
Note there was a system test which tested both the callout and the form
when unverified users visit a poll. We've split this system test in two
component tests.
Our original interface to vote in a poll had a few issues:
* Since there was no button to send the form, it wasn't clear that
selecting an option would automatically store it in the database.
* The interface was almost identical for single-choice questions and
multiple-choice questions, which made it hard to know which type of
question we were answering.
* Adding other type of questions, like open answers, was hard since we
would have to add a different submit button for each answer.
So we're now using radio buttons for single-choice questions and
checkboxes for multiple-choice questions, which are the native controls
designed for these purposes, and a button to send the whole form.
Since we don't have a database table for poll ballots like we have for
budget ballots, we're adding a new `Poll::WebVote` model to manage poll
ballots. We're using WebVote instead of Ballot or Vote because they
could be mistaken with other vote classes.
Note that browsers don't allow removing answers with radio buttons, so
once somebody has voted in a single-choice question, they can't remove
the vote unless they manually edit their HTML. This is the same behavior
we had before commit 7df0e9a96.
As mentioned in c2010f975, we're now adding the `ChangeByZero` rubocop
rule, since we've removed the test that used `and change`.
Until now, when writing `create(:poll_answer, option: option)`, the
`answer` field would take the title of a random option instead of taking
the title from the `option` variable.
So now, if we're given the option, the `answer` field will be taken from
the option itself.
Note that writing something like:
```
option { question.question_options.find_by(title: answer) }
answer { option.title }
```
Would create an infinite loop when creating a poll answer if we don't
pass the `option` and/or the `answer` attributes.
So, instead, we're making the `option` depend on the `answer` attribute
exclusively when we pass the `answer` attribute to the factory.
We used `margin-left` in commit b4eba055c, but when using right-to-left
layout, the property we should use is `margin-right`. So we're using
`margin-#{$global-left}` as usual.
