From: Phlip <phlipcpp@...>
Date: 2006-09-05T14:40:06+09:00
Subject: Re: run only one test case?

[One more irritation here - not all messages to Google Groups or the mailing 
list get thru to USENET...]

Logan Capaldo wrote:

> Now I see your problem.

You have a typo there; one extra 'y'.

>   def test_blah
>   end
>
> is NOT a test case as far as the established terminology is  concerned. It 
> is a test.

I once worked a project where the QA group insisted on saying "we regressed 
the current version today", when they meant, "We tested the current version 
against regressions today". Citing the dictionary definition of "regress" 
did not help them change their ways; apparently their industry has a streak 
of mis-use for that term.

    http://en.wikipedia.org/wiki/Test_case

"In software engineering, a test case is a set of conditions or variables 
under which a tester will determine if a requirement upon an application is 
partially or fully satisfied."

Naturally, that's the definition for manual testing. WikiPedia is free of 
points of view. The overall gist is that one (1) test case represents the 
Assemble Activate Assert cycle of one def test_blah(). An automated test 
case, within a version control system, will automatically satisfy each 
line-item of their manual test write-up. Note especially the line "Written 
test cases are usually collected into test suites."

There is a special reason most *Unit rigs call the parent of their suites 
'TestCase's. They implement the Interpreter Design Pattern, such that each 
Case may instead be a Suite containing many Cases, recursively. So the top 
thing's a Case, even though we normally derive from it to immediately create 
a Suite. The English connotations are a Case is a single unit and a Suite is 
a group of units.

AutoRunner noted; I will try it after I ... calm down. ;-) Thanks all.

-- 
  Phlip
  http://c2.com/cgi/wiki?ZeekLand <-- NOT a blog!!!