However, I didn't realize just how strong my addiction to tests had become! After a few hours of coding I couldn't keep going. The code felt wrong. I couldn't see any problems but I knew they must be there. I hadn't proved the code worked, so why would it? I started getting paranoid and my code got more and more defensive. Clearly this wasn't going to work.
So I broke down, added a t/ directory with a single test file and started back into test-driven development. The newish prove command made this particularly easy since I didn't need to add a Makefile, just:
$ prove t
Ah, much better now!
-sam
prove! (Score:1)
Re:prove! (Score:2)
http://www.petdance.com/perl/use-prove-lt.pdf [petdance.com]
Maybe I should distribute it with Test::Harness!
--
xoa
Same experience (Score:1)
What a nightmare! After about half a day, I was back writing tests again and I finally got productive. It truly is a myth that automated tests take more time.
MKDoc? (Score:1)
Re:MKDoc? (Score:2)
As to what MKDoc offers, I'm not entirely up-to-speed. It's a dynamic application platform, which makes it a very different kind of content-management system than Krang or Bricolage. It supports a lot of front-end interactivity that's hard to do with a static publisher: personalization, forums, polls, search, etc. There's also a lot of emphasis on accessability and