As mentioned in commits like a586ba806, a7664ad81, 006128da5, b41fbfa52
and c480cdd91, accessing the database after starting the browser with
the `visit` method sometimes results in database corruption and failing
tests on our CI due to the process running the test accessing the
database after the process running the browser has started.
In this case, we were hiding a proposal after starting the process
running the browser to check what happens when accessing a notification
for a hidden proposal. We can avoid database access in the middle of the
test by hidding a proposal before starting the browser. The process to
create a notification using the browser is already tested in other
specs, so we don't need to do it here as well.
Note that, to simplify the test, we're extracting the `notify_users`
method. I wonder whether this method should be called in an
`after_create` callback instead... That's a topic for another time,
though.
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.
JavaScript is used by about 98% of web users, so by testing without it
enabled, we're only testing that the application works for a very
reduced number of users.
We proceeded this way in the past because CONSUL started using Rails 4.2
and truncating the database between JavaScript tests with database
cleaner, which made these tests terribly slow.
When we upgraded to Rails 5.1 and introduced system tests, we started
using database transactions in JavaScript tests, making these tests much
faster. So now we can use JavaScript tests everywhere without critically
slowing down our test suite.
There seems to be a bug in the dashboard when calculating
`accumulat_supports` because in some cases `vote_weight` is `nil`.
Besides, these tests check the database after the browser has started,
instead of checking things from a user's point of view.
Since both the tests and the code are wrong, and I'm not familiar enough
with the code in the dashboard section, for now I'm leaving things as
they were.
We were displaying the total number of notifications with a message "You
have N unread notifications", but were using the total number of
notifications instead of the unread ones.
Other than simplifying the view, we can write tests using `click_link`,
which makes the tests more robust. Clicking the `.icon-notification`
element was causing some tests to fail when defining a window height of
750 pixels in the `admin_budgets` branch.
While running our test suite, we were getting an exception sometimes:
```
Proposal Notifications In-app notifications from the proposal's author Followers should receive a notification
Failure/Error: notification_for_user2 = Notification.find_by(user: user2)
ActiveRecord::StatementInvalid:
PG::ProtocolViolation: ERROR: bind message supplies 0 parameters, but prepared statement "" requires 2
: SELECT "notifications".* FROM "notifications" WHERE "notifications"."user_id" = $1 LIMIT $2
# ./spec/system/proposal_notifications_spec.rb:268
```
Sometimes we were getting a similar exception in a different test:
```
Commenting legislation questions Merged comment threads Reply on a multiple annotation thread and display it in the single annotation thread
And sometimes we were getting a different one:
Failure/Error: annotation.comments.roots.sort_by_most_voted.limit(Legislation::Annotation::COMMENTS_PAGE_SIZE).each do |comment|
ActionView::Template::Error:
PG::ProtocolViolation: ERROR: bind message supplies 0 parameters, but prepared statement "" requires 3
```
My best (wild) guess is these exceptions might take place because the
test is accessing the database and at the same time the browser
(chromedriver) process is accessing the database, with code like:
```
find(".icon-notification").click
notification_for_user2 = Notification.find_by(user: user2)
```
Or:
```
first(:css, ".annotator-hl").click
(...)
comment = annotation1.comments.first
click_link "Reply"
```
This behavior happened sometimes while using transactional fixtures and
a shared database connection with feature specs (before system specs
were introduced in Rails 5.1) when some queries were triggered from the
test after the browser process was started.
So we're avoiding the situation by writing the tests from the user's
point of view. This is just an attempt at fixing the issue; I don't know
whether these changes will fix it since I've only seen this exception on
Github Actions (never on my machine). Worst case scenario, we're still
improving the tests legibililty.
Using `travel` we go to `Time.now + interval`. If the application's time
zone changes due to seasonal clock changes during that interval, we
might travel to a time which is not the time we intended to travel to.
Example:
On a system using the UTC time zone, it's 9AM on October 25 (Friday).
Since by default CONSUL uses the Madrid time zone, it means the
application's time is 11AM.
We use `travel` to advance three days. That means we go to 9AM UTC on
October 28 (Monday). The application's time will be 10AM due to the
seasonal clock change, so we haven't fully traveled three days in
application's time.
To reproduce locally, run:
```
TZ=UTC rspec spec/system/proposal_notifications_spec.rb:410
```
Using `travel_to` with `3.days.from_now`, which uses the application's
time zone and so it will travel to October 28 at 11AM, solves the
problem.
