Re: RFC: GDL2 differences to EOF / WO4.5.1

[Giulio Cesare Solaroli]

On Thursday, Mar 27, 2003, at 09:43 Europe/Rome, David Ayers wrote:

> Giulio Cesare Solaroli wrote:
>
> > On Wednesday, Mar 26, 2003, at 22:53 Europe/Rome, Stéphane Corthésy 
> > wrote:
> >
> > > GDL2 won't have enough test cases, despite David's efforts, so if we 
> > > can get some more tests using OCUnit, that would benefit all.
> Yes, this is absolutly correct.  Right now the test cases merely cover 
> single classes and thier API and currently only include some EOControl 
> classes (Well also some adaptor and EOModel tests, but they need to be 
> revisited).  Also a lot of the tests are currently still "out of 
> context".  This should change over time, but for a more exhaustive 
> testing, additional testing frameworks such as yours can only help us.
>
> If I remember correctly, the only issue with OCUnit based code is 
> that, despite of that fact that it is based on free (as in speech) 
> software, it is not a GNU package, and that was why base our testing 
> on guile eventhough it is sometimes rather akward to handle.
>
> I think it would be great if you would contribute your extensions to 
> the FSF and you could maintain it in the GNUstep repository though (if 
> I interpreted you correctly).

I have no problem at all to contribute the code to FSF, but there are 
some issues that must be resolved:
- our frameworks use some of the Omni frameworks; before committing 
everything to the FSF this dependency must be removed; this should not 
be a huge problem, but must be investigated;
- some code snips have been taken from mailing-list messages; I have 
asked the original authors an informal permission when we did the first 
submission to SourceForge, but I think an further check should be done 
before giving the copyright to FSF (all the authors are cited in the 
source files);
- we subclass OCUnit classes so Sente too must contribute its code to 
FSF for complete integration.

This should be all that's needed.

By the way, inside our frameworks there is also a prototype for a 
Log4OC implementation; Mirko Viviani has done it upon our request. It's 
still a prototype, but I think somebody could be interested in it too. 
If Mirko agrees, this would be readily available for inclusion in the 
GNUStep project as it only rely on libxml.

Giulio Cesare