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

[Yen-Ju Chen]

> From: "Chris B. Vetter" <chrisv@web4inc.com>
> To: "Yen-Ju Chen" <yjchenx@hotmail.com>
> CC: discuss-gnustep@gnu.org
> Subject: Re: [Suggestion] GNUstep-test for quality control (WAS: Re: 
> deferred deallocation)
> Date: Tue, 14 Oct 2003 16:48:24 -0700
 [snip]
> > I know there are some testing code in GNUstep, and some of them use
> > guile which I will never try to use. Something like "gmake test" would
> > be more easy for normal user. "gmake test" is used in some of the
> > library.
>
> A _normal_ user doesn't know what a particular class is supposed to do,
> therefor a 'gmake test' won't do him/her much good, because he has no
> idea whether the result is correct or not.
>
> Besides that, a test scenario can only check so many features but not
> every possible issue.

 In general, "gmake test' might just report something like this:

 testing NSString.... pass
 testing NSNumber... pass
 testing NSBundle... fail (or more information)

 A well-written test suit can be used for each release
 to make sure nothing is broken when some other bugs are fixed.
 Users don't know the result. They just need to report when something fail
 and it will help the developers to discover the bugs.
 The goal of gnustep-test is to help discover the potential bugs
 and is easy for everyone to run.
 How well the gnustep-test can detect the bugs depends on how well it is 
written.
 We can just start some and add more when new bug is discovered.
 Over the time, it will cover over every details.

 Typo is another issue and I prefer not to touch this one. :)

 Yen-Ju

> --
> Chris

_________________________________________________________________
Add MSN 8 Internet Software to your current Internet access and enjoy 
patented spam control and more.  Get two months FREE!     
http://join.msn.com/?page=dept/byoa