I am trying to do the following in a Test/Unit integration test using Capybara. Very basic Rails 3 app, Ruby 1.9.2 just with the pg and capybara gems, thats it.
Everything works except that the Person object I create manually in the test does not show in the test database… or any other database I can find! But Person.all returns him, so he is getting created somewhere. Then, when Capybara/Selenium go to the browser which I have verified does hit the test database fine (through another test), I get an empty list of people on the index page.
Capybara/Selenium have the Rails.env correct in terms of hitting the db correctly. The Test itself says Rails.env = “test”, the data does not go into the test db. Where is the data getting saved? And how can I manually set up data in an Integration test before running Capybara? I thought maybe that Capybara was somehow triggering the database being emptied but the sequence below proves this is not the case:
def test_see_yoohoo
person = Person.new
person.name = “Yoohoo”
person.email = “yoohoo@gmail.com”
assert person.save # passes
assert Person.find_by_name(“Yoohoo”) # passes but in dbconsole “select * from people” in the test database (or dev database) returns no rows
assert_equal Rails.env, “test” # passes
visit "/people" # page loads fine.... Capybara/Selenium working....
assert page.has_content?('Yoohoo') # fails (page has empty list.... the list passes fine on another page so issue is not the view)
assert page.has_content?('yoohoo@gmail.com')
I am trying to do the following in a Test/Unit integration test using Capybara. Very basic Rails 3 app, Ruby 1.9.2 just with the pg and capybara gems, thats it.
Everything works except that the Person object I create manually in the test does not show in the test database… or any other database I can find! But Person.all returns him, so he is getting created somewhere. Then, when Capybara/Selenium go to the browser which I have verified does hit the test database fine (through another test), I get an empty list of people on the index page.
Capybara/Selenium have the Rails.env correct in terms of hitting the db correctly. The Test itself says Rails.env = “test”, the data does not go into the test db. Where is the data getting saved? And how can I manually set up data in an Integration test before running Capybara? I thought maybe that Capybara was somehow triggering the database being emptied but the sequence below proves this is not the case:
assert person.save # passes
assert Person.find_by_name("Yoohoo") # passes but in dbconsole "select * from people" in the test database (or dev database) returns no rows
assert_equal Rails.env, "test" # passes
visit "/people" # page loads fine.... Capybara/Selenium working....
assert page.has_content?('Yoohoo') # fails (page has empty list.... the list passes fine on another page so issue is not the view)
assert page.has_content?('yoohoo@gmail.com')
end
TestUnit methods are wrapped in a transaction, with every transaction being rollbacked after the method finishes.
You can set use_transactional_fixtures = false if you want to change this behavior.
Thanks! At first I did not believe this was the issue but I tried and it worked. So what I learned is that even though my calls to Capybara are within a single method the transactional fixture object is its own world. So it seems the ‘transaction’ never hits the db, right, so the tests are faster?
And so when the browser boots this is hitting a new transaction… and thus the data I just created is gone.
I am trying to do the following in a Test/Unit integration test using
Capybara. Very basic Rails 3 app, Ruby 1.9.2 just with the pg and
capybara
gems, thats it.
[...]
Is there any particular reason that you are torturing yourself this way
instead of using Cucumber?
