[rspec-users] Using RSpec to drive the design of a GUI desktop application

Eloy Duran eloy.de.enige at gmail.com
Sun Nov 18 07:36:10 EST 2007


Hey Mathijs,

Someone pinged me that you were interested in BDD with RubyCocoa (RC),
so I thought I'd just chip in.

I write my RC apps in a BDD way and as such support for it is included  
with Rucola.
When we release 0.0.2 I will also release a screencast on how I use  
BDD with RC.

But the bottom-line is, like with so many other BDD code, to make  
extensive use of mocking.
Because most of the times you don't want to test IF a NSTextField works,
you just want to test the behaviour of your code IF a text field  
returns a specific value.
This is easily done with mocking.

Atm I don't use rSpec but rather test/spec.
So there are probably some hurdles to overcome if you really wish to  
use rSpec.

Hope this helps.

Cheers,
Eloy Duran

===============================================================
Hello everybody,
I've been using RSpec as a tool to create web applications for some time
now, in Rails, and using plain Ruby with WEBrick as well. The tool  
suits my
needs and the story runner is great.
Now there are things that aren't solvable on the web, you'll need a  
_real_
desktop application for those problems.

So I've toyed a bit around with various GUI libraries as wxRuby and
RubyCocoa, to get a feeling on how these libraries work and I love to  
create
native OS X applications using cocoa.
Of course, the next question that arose in my head was:"How do I drive  
the
design of an application using a BDD framework like RSpec?".

When writing a web application, it is relatively easy to simulate a HTTP
request to the app and crawl through the returned HTML, but for a  
desktop
application it's different, right?
The format (html) handled between the application, and the toolkit  
drawing
the actual screen isn't that open for a desktop application as it is  
for a
web application.

So, since we don't want to test the inner workings of the gui toolkit  
and we
only want to specify the behaviour of the code we write self, we must  
plug a
framework somewhere, to capture the actual calls to this toolkit to  
know if
our code is doing the right thing. At least, that's how I see it  
currently.
But how could that be done?


Regards,

Matthijs Langenberg


More information about the rspec-users mailing list