I'm using Cucumber on my current project and it's causing me to wonder what other people's experience has been. I've read discussions about speeding up Cucumber and a basically everything else I can. Also, the rSpec book is articulate on the subject. What I'm curious about is whether a survey like this (in your opinions) would add useful information: Survey: How You Use Cucumber Do you: - Write failing steps first, then make them pass? - Use mocks/stubs? - Hit the database exactly as the application would or do you cache intermediate results? - Run: - Single scenarios instead of a whole feature - Individual features instead of single scenarios - All features instead of individual scenarios - Negative as well as positive scenarios - Use the helpful snippets for pending steps - Still use rSpec controller or view specs - Write features, then refactor parts out into shared steps later - Use the Textmate bundle - Run Cucumber all the time when developing controllers/views - Only run Cucumber for integration tests - Use an imperative or declarative style to describe scenarios - Ever include substitutable elements for readability that aren't used (i.e., /"(.*)" looks at his watch/ do |user| # user not referenced in step I'm happy to put a survey out there on whatever survey site people think is good once I have a good set of questions if there is interest and then tabulate the results. Good idea? Repetitive? Comments? Thanks, Steve
on 2009-02-03 19:44
on 2009-02-03 20:40
On Tue, Feb 3, 2009 at 9:42 AM, s.ross <email@example.com> wrote: > - Use the Textmate bundle > I just wanted to reply to this - Tim Pope's Cucumber plugin for Vim is awesome. You can be editing a feature step and go to its definition with a keystroke. ///ark
on 2009-11-05 11:49
Hi, So late reply after your post. I just started with cucumber and actually the list of questions you collected and answers to them (or points of views) would be quite nice. Even though different cases will require different approaches. But well if your idea ever materializes it will be interesting to check in my opinion. Of course participation is the key in a survey... Cheers.