Fix transient tests failures

[Rodrigora]

These changes fix the issue #961.

I added the timecop gem as suggested by @kurko.

Freezing the time in the two tests should fix the transient failures.

[bf4]

IMHO, any gem required for test to run should be in the gemspec as a development_dependency. your thoughts?

[bf4]

It too bad we can't just use https://github.com/rails/rails/blob/4-1-stable/activesupport/lib/active_support/testing/time_helpers.rb since they aren't in Rails 4.0

[Rodrigora]

@bf4 Ok. You're right. I'll move to gemspec file. I've found this article after your comment: http://yehudakatz.com/2010/12/16/clarifying-the-roles-of-the-gemspec-and-gemfile/

Another option is remove timecop dependency and use fixed datetimes like: Date.new(2015,6,6). What do you think?

[bf4]

I don't think that would work because it wouldn't affect the updated_at timestamps that Rails adds on save. You'd have to call model.update_column(:updated_at, Time.new(2015,6,6)) after the object was created ( see https://github.com/rails/rails/blob/4-2-stable/activerecord/lib/active_record/persistence.rb#L271-L291 ) or whatever the id is in this case

[Rodrigora]

Ok, but in these tests we don't use ActiveRecord inherited classes in tests.
The updated_at method is created "by hand" : https://github.com/rails-api/active_model_serializers/blob/master/test%2Ffixtures%2Fporo.rb#L21

And, the problem here is that it assumes that all code is executed in the same second. That's why sometimes it fails.

[bf4]

Good call

[bf4]

Thanks!

[bf4]

@kurko Any thoughts on this PR?

[kurko]

yup, I had suggested it to avoid time related issues.

[kurko]

Do you have concerns about timecop?

[joaomdmoura]

LGTM, @Rodrigora could you please squash the commits and rebase it? 😁

[bf4]

@joaomdmoura Looks like @Rodrigora squashed 'em

[Rodrigora]

@joaomdmoura, done. 👍

[joaomdmoura]

Great team work here ppl! Congratz! I'm merging it.