> What is the convention for naming and creating specs when speccing out cross
> object behaviour?
I missed this. For stuff like mixins, I'll generally stick the module
definition under lib/ and have a corresponding spec under spec/lib.
Also I usually write the spec like
module NameableSpec
class Person
include Nameable
end
describe Nameable do
before(:each) do
@person = Person.new
@person.set_name "Pat Maddox"
end
it "should parse the first name" do
@person.first_name.should == "Pat"
end
it "should parse the last name" do
@person.last_name.should == "Maddox"
end
end
end
The reason that I do that is because in order to write a spec for this
module, I have to actually mix it into a class. I wrap the whole spec
in a module so that my test class name doesn't clash with any existing
class names.
Another way to do it is to define the class dynamically before each test:
describe Nameable do
before(:each) do
klass = Class.new { include Nameable }
@person = klass.new
@person.set_name "Pat Maddox"
end
it "should parse the first name" do
@person.first_name.should == "Pat"
end
it "should parse the last name" do
@person.last_name.should == "Maddox"
end
end
The upside to that is that you don't have the ugly wrapper module.
The downside is that you can't easily reuse the class between describe
blocks - you'd still want to have a wrapper module to avoid polluting
the namespace. I tend to prefer this dynamic class creation, because
it's usually less code. I also end up with a bunch of little,
temporary helper classes that have only the behavior I want, rather
than a bigger test helper class with behavior to support multiple
specs.
I'm curious how other people handle this type of thing.
Pat
_______________________________________________
rspec-users mailing list
[email protected]
http://rubyforge.org/mailman/listinfo/rspec-users
