# Broken Tests in ActionPack

**URL:** <https://www.ruby-forum.com/t/broken-tests-in-actionpack/108477>\
**Category:** Rails\
**Created:** [July 3, 2007, 11:23pm UTC](https://www.ruby-forum.com/t/broken-tests-in-actionpack/108477 "2007-07-03T23:23:40Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![nickretallack](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/85e7bf/32.png) [@nickretallack](https://www.ruby-forum.com/u/nickretallack)\
**Post date:** [July 3, 2007, 11:23pm UTC](https://www.ruby-forum.com/t/broken-tests-in-actionpack/108477/1 "2007-07-03T23:23:40Z")

</div>

The ActionPack tests are borked. Unrelated tests affect each other.  
For proof, try removing the “unless defined?” clause around Customer  
in the beginning of actionpack/tests/controllers/render\_test.rb. It  
is actually using the incompatible definition from new\_render\_test.rb  
because of the order in which the tests are performed. However, the  
order is different on other platforms, so the “unless defined?” is  
really only hiding the problem from most users.

I hate to be a bother, but this really needs attention. I’d gladly  
fix it if someone were to suggest how this particular issue should be  
fixed. It scares me that unrelated tests can influence each other’s  
results. These are either poorly-written test files that break some  
fundamental rules of testing, or the test suite itself should be fixed  
to prevent this sort of thing from happening. Yes, we need to test  
our testing.

Here’s my original ticket: [http://dev.rubyonrails.org/ticket/8849](http://dev.rubyonrails.org/ticket/8849)

Please discuss.