Re: Re^3: Test output: Interpret or pass/fail

I used to, this way gives me three advantages.

1. I tend to work more efficantly when I can stay on track -- If I have an idea how to improve something and get side tracked on fixing a bug that just poped in the code then I take a while to get back into the flow once I fix it. If my bugs were usually severe and requiring a complete redesign this would be moot.

2. If I do implement solution X, and in the middle of implementation I figure out that solution Y is better for whatever reason I have many stop points I can jump back to implement solution Y.

3. I can audit my code almost change by change and show movment.

-Waswas

Comment on Re: Re^3: Test output: Interpret or pass/fail

Replies are listed 'Best First'.
Re^5: Test output: Interpret or pass/fail by adrianh (Chancellor) on Aug 07, 2003 at 06:55 UTC
I write test first and am very careful not to wear more than one of my programming, refactoring or debugging hats at any one time - so for me test-on-save works very well. As for stop points, code auditing, etc. I find a full source control system like CVS or Subversion a better fit than directory snapshots myself... but whatever works :-)	[reply]
Re^6: Test output: Interpret or pass/fail by waswas-fng (Curate) on Aug 07, 2003 at 07:03 UTC
I use the snapshots for minors and CVS nightly once I finish up what I am doing. I don't need 5000 versions in cvs, I delete the snapshot tree weekly after I am sure about the direction I have taken it. -Waswas	[reply]