package EOL

[Patrice Dumas]

> > I personally think that it is acceptable and shouldn't be a reason not 
> > to ship those programs
> 
> I didn't propose to stop shipping them, but putting them in a separate
> Fedora repo to highlight the fact the level of support/testing/fixing
> which can be expected is way lower than that of main FE repo itself.

Ok, why not if it's not complicated and don't add trouble, but I think
that packager should not be forced to put them in (orphans could go in
'automatically').

> It's way better to have a gray repo rather than putting black sheep in
> the main repo just because some packagers find them too convenient to
> drop (and are ready to accept the convenience/security compromise a
> uninformed user may not find so cool)

Ok.

> If the lib is useful then there should be at least one packageable user.
> If *no* lib user can be packaged :
> 1. then it may not be so useful after all
> 2. or it's only used by unpackageable stuff, in which case I  don't see
> why it should be packaged when its users aren't

Once more I'll take the example of the cernlib. In the cernlib there are
many usefull math subroutines that may be used in numerical models, though
I doubt anything will ever depend on the cernlib. I doubt a model will 
ever go in fedora...

There is no general case, however, and I am not against arguing your point
for a particular lib.

> Also it's going to be loads of fun when users report problems with the
> lib and there are no users in the repo to test it against/reproduce the
> problem

Not to worry about. If a user wants to reproduce something he will 
provide the code.
 
> If it's that obvious to you I don't see what's wrong in writing it.
> Consensus - yep

Ok for that one... But that makes 3 out of 6...

--
Pat