Re: [Suggestion] GNUstep-test for quality control (WAS: Re: deferreddeallocation)

[Chris B. Vetter]

On Fri, 17 Oct 2003 21:08:33 +0200
Alexander Malmberg <alexander@malmberg.org> wrote:
[...]
> Chris Hanson wrote:
> > On Tuesday, October 14, 2003, at 06:48  PM, Chris B. Vetter wrote:
> > > Besides that, a test scenario can only check so many features but
> > > not every possible issue.
> > That's not necessarily the case.  You can get very high test
> > coverage if you develop code test-first using a good unit testing
> > framework, Test Driven Development, and refactoring techniques.
> Well, that depends on what you want to cover. Generally, I'd say that
> test-first unit tests can achieve good coverage of features of the
> "interface", while bug-based tests cover features of the
> "implementation".
> Anyway, the merits of test-first are fairly irrevelant in our case
> since most of the code that needs testing has already been written. :)
[...]

Still, I don't see how this will cover 'faulty' implementations. Yes,
you will be able to tell whether it works. Yes, you will be able to find
(at least some) bugs. But no, you won't be able to tell, whether the
current implementation is _correct_ that is, follows the specification.

As an example (that might be a bit far-fetched, I admit):

I was looking at Chad's GSDock and noticed that a double click on a
'swallowed' application tile will pop open another application tile at
the bottom of your screen.

Now, the OpenStep specification, for this particular case, mentions a
method in NSWorkspace: -launchApplication:showTile:autolaunch:

  [quote]
  Instructs Workspace Manager to launch the application appName.  If
  showTile is NO, Workspace Manager won't display a tile for the
  application.  (The tile will exist, but it won't be placed on the
  screen.) [...]
  [end quote]

GNUstep's NSWorkspace offers -launchApplication:showIcon:autolaunch:
which is, as to my understanding, the implementation of above mentioned
method.

GNUstep DOES launch the application, however it does NOT suppress the
icon. Which is, IMHO, a bug in the current implementation.

There's no way a test unit will be able to cover these issues.

-- 
Chris