views:

70

answers:

2

I'm trying to research a term for software testing. Specifically, that applies to the following scenario:

  1. You've got software "S", which is in version V
  2. S has functionality to "whiz," "bang," and "zoom"
  3. You find a bug with S's zooming
  4. You apply a patch to address the zoom function
  5. S can now zoom
  6. S can no longer whiz.

My question: What types of tests are specifically designed to ensure that a patch to a particular module doesn't unexpectedly break other modules? Is this simply a special case of Regression Testing? is there a special term for this?

Apologies if this is a repost, and thanks in advance!

Cheers

+4  A: 

Yep, that kind of bug is called a regression. And regression testing is a basic component of (automated and manual) testing. This class of tests often just arises from creating tests for new stuff and specific bugs along the way, but continuing to verify new builds against the whole cumulative test set.

quixoto
Awesome - just the explanation I was looking for. Thanks!
Chris
A: 

I sometimes call this phenomenon "Whack-a-Mole."

You fix one thing, and another problem pops up somewhere else.

If your modules are unit testable, you can try to write as many tests as possible to cover potential scenarios like this. If the modules can't be realistically unit tested, then yeah, you'll have to regression test the whole system to catch things like this.

Andy White