Remove warning caused by changes in minitest 5

[vanstee]

Fixes the warning reading: "Warning: you should require
'minitest/autorun' instead."

Minitest 5 moved a few things around, such as the "MiniTest" to
"Minitest" module change, so those were updated as well.

[vanstee]

Ran into this running Rails 4.1.0.beta and Ruby 2.0.0-p195. This is also probably a fix for #758

[fables-tales]

@vanstee this mostly looks good to me. @alindeman will probably have opinions. For my 2c I think this needs to go into 2-14, 2-99 and 3. Andy do you agree?

[vanstee]

D'oh. This blows up on older ruby versions. Maybe I can wrap this stuff up nicely and switch depending on the ruby version. Thoughts?

[fables-tales]

@vanstee however it works we need to pass under 1.8.7. if you've got thoughts on how to do it then by all means go ahead!

[alindeman]

Is this ready to be reviewed again? Seems like there are still some Travis errors.

Thanks for looking into this, @vanstee.

[vanstee]

Oops. Sorry about not following up on this.

I was hoping we could just require minitest ~> 5.0 but that also seems to be throwing a few errors. I'll try to spend some time on it this week.

[vanstee]

Well that's a start. Tests pass until you hit the same problem in cucumber. I'll clean some stuff up later and get this ready for merging.

[vanstee]

Yay! All the test failures seem to be unrelated and I'm no longer seeing the warning on master. @alindeman let me know if there's anything else I can do to clean this up.

[vanstee]

Oops I guess I should probably leave the require up at the top. Maybe this should all be hidden away in a separate file as well.

[alindeman]

Oops I guess I should probably leave the require up at the top. Maybe this should all be hidden away in a separate file as well.

Is this comment still relevant? If so, I need a bit more info on what you mean.

[alindeman]

Other than the one error caused by a rubygems.org timeout, any reason not to trust the all green here and merge? https://travis-ci.org/alindeman/rspec-rails/builds/9561178

I believe the one failure on master is Rails' fault. I might look into it sometime.

[vanstee]

@alindeman I split out each adapter to clean things up. Let me know if the extra requires are something we should worry about. I can always remove those last two commits.

[bricker]

This pull request fixes more than just the warnings. Loading test/unit/assertions with minitest > 5.0 and ruby 1.9.3-p0 causes uninitialized constant Minitest::MINI_DIR. See rails/rails#11691 for more details.

Here is an example app to reproduce the failure: https://github.com/bricker/ruby-1.9.3-p0-failure

A new error occurrs when running the example app's specs with this patch (plus upgrading to rspec 3.0.0.pre), but it's unrelated. The uninitialized constant error goes away.

[alindeman]

I'm playing with this a bit, and I think it's close. I'll plan to have it merged in by the weekend, barring any big issues.

Thanks so much for getting this together @vanstee.

[vanstee]

@alindeman Thanks for the review. Let me know if there's anything I can do to clean it up. Wasn't sure if splitting up those adapter classes for readability was worth the extra requires or not.

[alindeman]

@alindeman Thanks for the review. Let me know if there's anything I can do to clean it up. Wasn't sure if splitting up those adapter classes for readability was worth the extra requires or not.

For now, I think I may leave out that commit. It makes thing a bit harder to find, and we have to be sensitive of the number of requires we do. It's easy enough to take out, so no worries.

[alindeman]

Merged with f4dd001. I will likely go back to 2.99 and maybe 2.14 with this.

If you get a sec, @vanstee, do you mind looking over it and making sure it works for you?

[vanstee]

@alindeman Was only able to test with 2-14-maintenance due to some dependencies in other gems. Looks good 👍

[gjaldon]

Experienced this same problem with Rails-4.1.0.beta. Just upgraded to it and am using Ruby 2.0.0p353. Wasn't experiencing this issue before the upgrade. However, if I require 'rspec/autorun' before require 'rspec/rails', the error/warning is silenced.

[vanstee]

@gjaldon Looks like @alindeman fixed this one but I don't think it's been released as part of 2.14.x yet. Although, you could give 2.99.0.beta1 or 3.0.0.beta1 a try. I've been using 2.99.0.beta1 for a couple of weeks without any issues.

[gjaldon]

@vanstee thanks for the info. Will try that out. Btw, the warning isn't silenced when changing the order of requires. Wrong statement. :)

[noahsw]

I get this warning only when I have gem 'shoulda-matchers' in my Gemfile. I'm using rspec-rails 3.0.0.beta1 with rails 3.2.12. Changing the order of require 'rspec/autorun' and require 'rspec/rails' did not have an effect.

[pardel]

I can confirm @noahsw's scenario. It happens the moment I'm adding 'shoulda-matchers'.

[alindeman]

2.14.1 is released. Please open a new issue if there are still problems :)

[danielmurphy]

Upgrading to 2.14.1 fixed it for me. Thanks!

[noahsw]

Hmm... I've been using 2.14.1 this whole time. It still has the issue.

RubyGems Environment:
  - RUBYGEMS VERSION: 1.8.23
  - RUBY VERSION: 1.9.3 (2013-06-27 patchlevel 448) [x86_64-darwin12.4.0]

rails (3.2.12)
rspec (2.14.1)
rspec-collection_matchers (0.0.2)
rspec-core (3.0.0.beta1, 2.14.7)
rspec-expectations (3.0.0.beta1, 2.14.4)
rspec-mocks (3.0.0.beta1, 2.14.4)
rspec-rails (3.0.0.beta1)
rspec-support (3.0.0.beta1)
shoulda-matchers (2.4.0)

[emaiax]

I'm still experiencing this issue.

My Gemfile.lock looks like this.

rails (4.1.0.beta1)
factory_girl (4.3.0)
rake (10.1.1)
rspec-core (2.14.7)
rspec-expectations (2.14.4)
  diff-lcs (>= 1.1.3, < 2.0)
rspec-mocks (2.14.4)
rspec-rails (2.14.1)
  actionpack (>= 3.0)
  activemodel (>= 3.0)
  activesupport (>= 3.0)
  railties (>= 3.0)
  rspec-core (~> 2.14.0)
  rspec-expectations (~> 2.14.0)
  rspec-mocks (~> 2.14.0)

[zolzaya]

Same here :(

My environments:

Gems included by the bundle:
  * actionmailer (4.1.0.rc1)
  * actionpack (4.1.0.rc1)
  * actionview (4.1.0.rc1)
  * activemodel (4.1.0.rc1)
  * activerecord (4.1.0.rc1)
  * activesupport (4.1.0.rc1)
  * addressable (2.3.5)
  * admin (0.0.1)
  * arel (5.0.0)
  * atomic (1.1.14)
  * bcrypt-ruby (3.1.2)
  * bootstrap-sass (3.1.1.0)
  * builder (3.2.2)
  * bundler (1.3.5)
  * capistrano (3.1.0)
  * capistrano-bundler (1.1.2)
  * capistrano-rails (1.1.1)
  * capistrano-rvm (0.1.1)
  * capybara (2.2.1)
  * celluloid (0.15.2)
  * coderay (1.1.0)
  * coffee-rails (4.0.1)
  * coffee-script (2.2.0)
  * coffee-script-source (1.6.3)
  * database_cleaner (1.2.0)
  * delayed_job (4.0.0)
  * delayed_job_active_record (4.0.0)
  * diff-lcs (1.2.5)
  * docile (1.1.2)
  * em-websocket (0.5.0)
  * erubis (2.7.0)
  * eventmachine (1.0.3)
  * execjs (2.0.2)
  * factory_girl (4.4.0)
  * factory_girl_rails (4.4.0)
  * ffaker (1.23.0)
  * ffi (1.9.3)
  * formatador (0.2.4)
  * guard (2.3.0)
  * guard-livereload (2.1.1)
  * guard-rspec (4.2.2)
  * hike (1.2.3)
  * http_parser.rb (0.5.3)
  * i18n (0.6.9)
  * jbuilder (1.5.3)
  * jquery-rails (3.1.0)
  * json (1.8.1)
  * kaminari (0.15.1)
  * launchy (2.4.2)
  * listen (2.4.0)
  * lumberjack (1.0.4)
  * mail (2.5.4)
  * method_source (0.8.2)
  * mime-types (1.25.1)
  * mini_portile (0.5.2)
  * minitest (5.2.3)
  * multi_json (1.8.4)
  * net-scp (1.1.2)
  * net-ssh (2.8.0)
  * nokogiri (1.6.1)
  * pg (0.17.1)
  * polyglot (0.3.4)
  * pry (0.9.12.6)
  * rack (1.5.2)
  * rack-test (0.6.2)
  * rails (4.1.0.rc1)
  * railties (4.1.0.rc1)
  * rake (10.1.1)
  * rb-fsevent (0.9.4)
  * rb-inotify (0.9.3)
  * rspec (3.0.0.beta1)
  * rspec-collection_matchers (0.0.3)
  * rspec-core (3.0.0.beta1)
  * rspec-expectations (3.0.0.beta1)
  * rspec-mocks (3.0.0.beta1)
  * rspec-rails (3.0.0.beta1)
  * rspec-support (3.0.0.beta1)
  * sass (3.2.13)
  * sass-rails (4.0.1)
  * shoulda-matchers (2.5.0)
  * simplecov (0.8.2)
  * simplecov-html (0.8.0)
  * slop (3.4.7)
  * spring (1.1.1)
  * spring-commands-rspec (1.0.1)
  * sprockets (2.10.1)
  * sprockets-rails (2.0.1)
  * sshkit (1.3.0)
  * term-ansicolor (1.3.0)
  * thor (0.18.1)
  * thread_safe (0.1.3)
  * tilt (1.4.1)
  * timers (1.1.0)
  * tins (1.0.0)
  * treetop (1.4.15)
  * turbolinks (2.2.1)
  * tzinfo (1.1.0)
  * uglifier (2.4.0)
  * warden (1.2.3)
  * xpath (2.0.0)

[JonRowe]

@zolzaya @emaiax can one of you produce an isolated example of this? (Create a new empty rails app and use your current gemfile.lock, check it raises the error and then push it to a gh repo)

[zolzaya]

@JonRowe Okay give me a minute

[zolzaya]

@JonRowe OMG it's working without warning in new rails application. But I got same warning in my old application. Hmm ..................

[dpickett]

If you're still getting this error there's a chance it could be shoulda-matchers - if I add require: false to my shoulda reference in my Gemfile, the error is silenced.

[tuomasj]

Thanks @dpickett, it silences the error. I'm running Rails 4.1.0.rc1 and RSpec 3.0.0.beta2

gem 'shoulda', require: false

[allcentury]

Thanks @dpickett ! Worked for me as well.