We've had some tests fail after this test was executed in our CI, and
one possible reason could be that sometimes this test finished before
all its requests had finished. This could be the case with the following
code:
```
visit sdg_management_proposals_path(filter: "pending_sdg_review")
click_button "Search"
expect(page).to have_css "li.is-active h2", exact_text: "Pending"
```
Before clicking the "Search" button, the expectation is already true, so
there's a chance that the expectation is evaluated as true before the
request has finished. That might result in requests and session data
leaking between tests.
So we're adding more expectations in order to make sure that the
requests have finished before evaluating the expectations associated to
them.
All the code in the `bin/` and the `config/` folders has been generated
running `rake app:update`. The only exception is the code in
`config/application.rb` where we've excluded the engines that Rails 6.0
has added, since we don't use them.
There are a few changes in Active Storage which aren't compatible with
the code we were using until now.
Since the method to assign an attachment in ActiveStorage has changed
and is incompatible with the hack we used to allow assigning `nil`
attachments, and since ActiveStorage now supports assigning `nil`
attachments, we're removing the mentioned hack. This makes the
HasAttachment module redundant, so we're removing it.
Another change in ActiveStorage is files are no longer saved before
saving the `ActiveStorage::Attachment` record. This means we need to
manually upload the file when using direct uploads. We also have to
change the width and height validations we used for images; however,
doing so results in very complex code, and we currently have to write
that code for both images and site customization images.
So, for now, we're just uploading the file before checking its
dimensions. Not ideal, though. We might use active_storage_validations
in the future to fix this issue (when they support a proc/lambda, as
mentioned in commit 600f5c35e).
We also need to update a couple of tests due to a small change in
response headers. Now the content disposition returns something like:
```
attachment; filename="budget_investments.csv"; filename*=UTF-8''budget_investments.csv
```
So we're updating regular expression we use to check the filename.
Finally, Rails 6.0.1 changed the way the host is set in integration
tests [1] and so both `Capybara.app_host` and `Capybara.default_host`
were ignored when generating URLs in the relationable examples. The only
way I've found to make it work is to explicitely assign the host to the
integration session. Rails 6.1 will change this setup again, so maybe
then we can remove this hack.
[1] https://github.com/rails/rails/pull/36283/commits/fe00711e9
While this bug was already present in the general admin search, the
combination of both search and filters was very uncommon. I've only
found this combinations in the users section, where you've got the
"erased" filter, but in this case searching for erased users doesn't
really make sense since their username and email have been deleted and
so there's nothing to find.
So the hidden content seemed to be the only affected section. However,
we're adding the field to every section so we don't have to make sure we
add it when we need it (like we did in the SDGManagement section).
Rubocop was complaining about a Layout/ExtraSpacing in a couple of
places.
These issues weren't detected by Pronto because they didn't affect lines
changed in the pull request. These lines were fine until we removed the
lines next to them in commits 4b42a68b6 and 00f0c4410.
This test has been failing many times because it checks the database
after starting the process running the browser.
We're disabling JavaScript like we do in every other test using the
`create_proposal_notification` method. This way, Capybara will use the
rack driver and there'll be no risk of errors that currently might take
place if both the process running the browser and the process running the
test access the database.
When we perform database queries in tests after the process running the
browser has started, we sometimes get failures in our test suite due to
both the tests and the browser accessing the database at the same time.
Furthermore, using `Poll.all` results in a database query, and doing so
after the process running the browser has started might result in
failures when running our test suite.
We were clicking on the "Sign in" link right after clicking on the "Sign
out" link, which might result in simultaneous requests and exceptions
when running our test suite.
So we're adding an expectation to make sure the first request has
finished before starting the following one.
We were missing a notice in this case. Not only this caused
inconsistencies in the user experience, but it also made it hard to add
an expectation in the test checking the request had finished before
making a new one. Simultaneous requests sometimes cause failures in our
test suite.
Currently with both seeds and dev_seeds, not only was this email not
displayed from the system emails section, but it also caused an error in
the application.
@email_to had an empty value and in the view we tried to access
@email_to.name which caused the error. We kept the same logic but
added the current_user to make sure it always has a valid value. We add
the current_user because the current_user is always present in this controller..
"Participate in the final voting" was a concept which was hard to
understand since many people would think it was related to the
voting/polls section and that somehow there was going to be a "final"
poll.
So we use "Vote for budget projects" instead.
Thanks Pomerange for the suggestion.
We had the same texts four times, with slight variations in the case of
the management section.
We're unifying them under the "verification" i18n namespace, since the
texts are about actions which can be done depending on whether users are
verified or not.
Note the names of the i18n keys aren't very consistent, since we use
"debates" in plural but "proposal" in singular. We're leaving it like
this so existing translations aren't affected.
We used "retire" because we translated it literally from the Spanish
verb "retirar" which can mean both "retire" and "withdraw".
Note we're still using "retire" in database fields and method names;
changing that might make it harder to upgrade from a previous version of
CONSUL.
On the Configuration settings page three settings appeared without
description:
* Comments Summary: No description.
* Related Content: No description.
* Tags: No description.
These settings are related with the AI / Machine learning feature. They
only should appear on AI / Machine learning setting page when the
feature is enabled.
It looks like these lines were added on a branch which didn't include
commit 3da4ee00b but were merged after that commit was merged.
In any case, since we're already using the `:admin` tag in the whole
file, these lines aren't necessary.
After a user assigned as a budget admin deletes their account or gets blocked by
a moderator, the application throws an exception while loading the admin
investment index page.
As an erased user is not really deleted and neither its associated roles, the
application was failing when trying to sort and administration without a
username. In this case, the application was throwing an `ArgumentError:
comparison of NilClass with String failed` exception.
As a blocked user is not deleted or its roles, the application failed when trying
to access the user name through the delegation in the Administrator. In this
case, the application was throwing a `NoMethodError: undefined method `name' for
nil:NilClass` exception.
While reviewing commit 7702b551e, I forgot to test whether selecting
text in the budget header or clicking its link worked properly.
The negative index (-5) meant it was impossible to select its text or
click on its link.
The good news is the pseudoelement with a negative index (-1) is
considered a child of the .budget-header element, so having a negative
index will cause the pseudoelement to be render behind the content of
the .budget-header element but in front of the background of the
.budget-header element.
This is exactly what we want.
Originally, we didn't have a z-index in the .budget-header element,
meaning the pseudoelement was rendered behind the background of the
.budget-header element, meaning both backgrounds were visible. This was
OK when the background was a plain color, but it wasn't when the
background was an image.
To stress the fact that the budget header is only affected when we use
an image, I'm also moving the code inside the `.with-background-image`
selector, although it would be interesting to check whether it's a good
idea to add `z-index: 0` to the `full-width-background` mixin.
We forgot to do so in commit 04605d5d5. Before that commit, we were
testing the links weren't displayed in the "informing" phase as opposed
to the "finished" phase. After that commit, they weren't displayed
anywhere since a <select> field was shown instead.
Since then, we've replaced the <select> field with the links generated
by the `Budgets::Investments::FiltersComponent`. We've already got tests
for these links.
We're going to make a change, and it's easier if we've already got a
component with tests so we don't have to write system tests to check
whether the map is rendered.
