Ruby require 'file' and relative location

Question

So I'm writing some rspec tests and I'm embarrassed at my lack of Ruby understanding.

I have a file structure that looks like the following:

• GUI_Tests/Tests/test_spec.rb
• GUI_Tests/windows_gui.rb
• GUI_Tests/upload_tool.rb

when I run spec for the test_spec.rb file, I require the upload_tool file to be included like so:

spec -r ../upload_tool -fs test_spec.rb

Then, the upload_tool requires windows_gui.rb, like so:

require '../windows_gui'

My question is, why so I have to reference windows_gui.rb relative to test_spec.rb (requiring the ../) rather than the upload_tool.rb? This feels wrong to me, I'll want to use the upload_tool.rb out of context of the test specs, which means changing the requires each time.

Clearly I'm missing something, but if I don't reference relative to the test spec I get a file not found error.

Sorry for being so ignorant here, but I'm coming up empty handed. Any thoughts appreciated.

BB

Scriptural

Don't feel bad - I find dealing with relative paths one of the most annoying things about ruby. I've listed it in what are the things you would like improved in the ruby language – Cranach

August 2012 - Note that this post is now two years old. Most folks are now on 1.9.2. See this post https://mcmap.net/q/667651/-relative-file-path-in-rspec also for the best solution (Feb 2012) – Isoprene

Answer 1

You don't. requires are relative to the current directory, which in your case was GUI_Tests/Tests. If you did this instead:

cd ..
spec -r upload_tool -fs Test/test_spec.rb

You would have to use this:

require 'windows_gui' # without '../'

The most common way to get around that problem is using File.dirname(__FILE__):

require File.join(File.dirname(__FILE__), '..', 'windows_gui')

NOTE: in Ruby 1.9.2 require changed it's defaults: Ruby: require vs require_relative - best practice to workaround running in both Ruby <1.9.2 and >=1.9.2

Staggard

Perfect - like I said, I was embarrassed for even having to ask, but I'm glad I did. That makes perfect sense. – Scriptural

Actually, require is not relative to the current directory, it is relative to the search path (i.e. the special constant $LOAD_PATH or its alias $: ). It just so happens that on older versions of Ruby (i.e. all except the not-yet-but-soon released 1.9.2), the current directory ( . ) is in the search path by default: p $:.include?('.') # => true . However, as of Ruby 1.9.2 that is no longer the case and shouldn't be relied upon. In Ruby 1.9, you could use require_relative which works relative to the source file it is called from instead of the search path. – Noonday

@Jorg: But if I run this in 1.9: $:.replace []; require '../somefile', it works. – Staggard